fix(backend): Include is_enabled in public model list response

Introduce a new boolean property `is_enabled` (default: true) into the OpenAPI schema in autogpt_platform/frontend/src/app/api/openapi.json next to `price_tier` and `is_recommended`. This exposes an enable/disable flag in the API model for consumers and defaults new entries to enabled.

autogpt_platform/backend/backend/api/rest_api.py

@@ -1,3 +1,4 @@
+import asyncio
 import contextlib
 import logging
 import platform
@@ -37,8 +38,10 @@ import backend.api.features.workspace.routes as workspace_routes
 import backend.data.block
 import backend.data.db
 import backend.data.graph
+import backend.data.llm_registry
 import backend.data.user
 import backend.integrations.webhooks.utils
+import backend.server.v2.llm
 import backend.util.service
 import backend.util.settings
 from backend.api.features.library.exceptions import (
@@ -117,16 +120,56 @@ async def lifespan_context(app: fastapi.FastAPI):
 
     AutoRegistry.patch_integrations()
 
+    # Load LLM registry before initializing blocks so blocks can use registry data.
+    # Tries Redis first (fast path on warm restart), falls back to DB.
+    # Note: Graceful fallback for now since no blocks consume registry yet (comes in PR #5)
+    try:
+        await backend.data.llm_registry.refresh_llm_registry()
+        logger.info("LLM registry loaded successfully at startup")
+    except Exception as e:
+        logger.warning(
+            f"Failed to load LLM registry at startup: {e}. "
+            "Blocks will initialize with empty registry."
+        )
+
+    # Start background task so this worker reloads its in-process cache whenever
+    # another worker (e.g. the admin API) refreshes the registry.
+    _registry_subscription_task = asyncio.create_task(
+        backend.data.llm_registry.subscribe_to_registry_refresh(
+            backend.data.llm_registry.refresh_llm_registry
+        )
+    )
+
     await backend.data.block.initialize_blocks()
 
     await backend.data.user.migrate_and_encrypt_user_integrations()
     await backend.data.graph.fix_llm_provider_credentials()
-    await backend.data.graph.migrate_llm_models(DEFAULT_LLM_MODEL)
+    try:
+        await backend.data.graph.migrate_llm_models(DEFAULT_LLM_MODEL)
+    except Exception as e:
+        err_str = str(e)
+        if "AgentNode" in err_str or "does not exist" in err_str:
+            logger.warning(
+                f"migrate_llm_models skipped: AgentNode table not found ({e}). "
+                "This is expected in test environments."
+            )
+        else:
+            logger.error(
+                f"migrate_llm_models failed unexpectedly: {e}",
+                exc_info=True,
+            )
+
     await backend.integrations.webhooks.utils.migrate_legacy_triggered_graphs()
 
     with launch_darkly_context():
         yield
 
+    _registry_subscription_task.cancel()
+    try:
+        await _registry_subscription_task
+    except asyncio.CancelledError:
+        pass
+
     try:
         await shutdown_cloud_storage_handler()
     except Exception as e:
@@ -355,6 +398,11 @@ app.include_router(
     tags=["oauth"],
     prefix="/api/oauth",
 )
+app.include_router(
+    backend.server.v2.llm.router,
+    tags=["v2", "llm"],
+    prefix="/api",
+)
 
 app.mount("/external-api", external_api)
 

autogpt_platform/backend/backend/data/graph.py

@@ -38,7 +38,7 @@ from backend.util.request import parse_url
 from .block import BlockInput
 from .db import BaseDbModel
 from .db import prisma as db
-from .db import query_raw_with_schema, transaction
+from .db import execute_raw_with_schema, query_raw_with_schema, transaction
 from .dynamic_fields import is_tool_pin, sanitize_pin_name
 from .includes import AGENT_GRAPH_INCLUDE, AGENT_NODE_INCLUDE, MAX_GRAPH_VERSIONS_FETCH
 from .model import CredentialsFieldInfo, CredentialsMetaInput, is_credentials_field_name
@@ -1669,16 +1669,15 @@ async def migrate_llm_models(migrate_to: LlmModel):
 
     # Update each block
     for id, path in llm_model_fields.items():
-        query = f"""
-            UPDATE platform."AgentNode"
+        query = """
+            UPDATE {schema_prefix}"AgentNode"
             SET "constantInput" = jsonb_set("constantInput", $1, to_jsonb($2), true)
             WHERE "agentBlockId" = $3
             AND "constantInput" ? ($4)::text
-            AND "constantInput"->>($4)::text NOT IN {escaped_enum_values}
-            """
+            AND "constantInput"->>($4)::text NOT IN """ + escaped_enum_values
 
-        await db.execute_raw(
-            query,  # type: ignore - is supposed to be LiteralString
+        await execute_raw_with_schema(
+            query,
             [path],
             migrate_to.value,
             id,

autogpt_platform/backend/backend/data/llm_registry/__init__.py

@@ -0,0 +1,40 @@
+"""LLM Registry - Dynamic model management system."""
+
+from backend.blocks.llm import ModelMetadata
+from .notifications import (
+    publish_registry_refresh_notification,
+    subscribe_to_registry_refresh,
+)
+from .registry import (
+    RegistryModel,
+    RegistryModelCost,
+    RegistryModelCreator,
+    clear_registry_cache,
+    get_all_model_slugs_for_validation,
+    get_all_models,
+    get_default_model_slug,
+    get_enabled_models,
+    get_model,
+    get_schema_options,
+    refresh_llm_registry,
+)
+
+__all__ = [
+    # Models
+    "ModelMetadata",
+    "RegistryModel",
+    "RegistryModelCost",
+    "RegistryModelCreator",
+    # Cache management
+    "clear_registry_cache",
+    "publish_registry_refresh_notification",
+    "subscribe_to_registry_refresh",
+    # Read functions
+    "refresh_llm_registry",
+    "get_model",
+    "get_all_models",
+    "get_enabled_models",
+    "get_schema_options",
+    "get_default_model_slug",
+    "get_all_model_slugs_for_validation",
+]

autogpt_platform/backend/backend/data/llm_registry/notifications.py

@@ -0,0 +1,84 @@
+"""Pub/sub notifications for LLM registry cross-process synchronisation."""
+
+from __future__ import annotations
+
+import asyncio
+import logging
+from typing import Awaitable, Callable
+
+logger = logging.getLogger(__name__)
+
+REGISTRY_REFRESH_CHANNEL = "llm_registry:refresh"
+
+
+async def publish_registry_refresh_notification() -> None:
+    """Publish a refresh signal so all other workers reload their in-process cache."""
+    from backend.data.redis_client import get_redis_async
+
+    try:
+        redis = await get_redis_async()
+        await redis.publish(REGISTRY_REFRESH_CHANNEL, "refresh")
+        logger.debug("Published LLM registry refresh notification")
+    except Exception as e:
+        logger.warning("Failed to publish registry refresh notification: %s", e)
+
+
+async def subscribe_to_registry_refresh(
+    on_refresh: Callable[[], Awaitable[None]],
+) -> None:
+    """Listen for registry refresh signals and call on_refresh each time one arrives.
+
+    Designed to run as a long-lived background asyncio.Task.  Automatically
+    reconnects if the Redis connection drops.
+
+    Args:
+        on_refresh: Async callable invoked on each refresh signal.
+                    Typically ``llm_registry.refresh_llm_registry``.
+    """
+    from backend.data.redis_client import HOST, PASSWORD, PORT
+    from redis.asyncio import Redis as AsyncRedis
+
+    while True:
+        try:
+            # Dedicated connection — pub/sub must not share a connection used
+            # for regular commands.
+            redis_sub = AsyncRedis(
+                host=HOST, port=PORT, password=PASSWORD, decode_responses=True
+            )
+            pubsub = redis_sub.pubsub()
+            await pubsub.subscribe(REGISTRY_REFRESH_CHANNEL)
+            logger.info("Subscribed to LLM registry refresh channel")
+
+            while True:
+                try:
+                    message = await pubsub.get_message(
+                        ignore_subscribe_messages=True, timeout=1.0
+                    )
+                    if (
+                        message
+                        and message["type"] == "message"
+                        and message["channel"] == REGISTRY_REFRESH_CHANNEL
+                    ):
+                        logger.debug("LLM registry refresh signal received")
+                        try:
+                            await on_refresh()
+                        except Exception as e:
+                            logger.error(
+                                "Error in registry on_refresh callback: %s", e
+                            )
+                except asyncio.CancelledError:
+                    raise
+                except Exception as e:
+                    logger.warning(
+                        "Error processing registry refresh message: %s", e
+                    )
+                    await asyncio.sleep(1)
+
+        except asyncio.CancelledError:
+            logger.info("LLM registry subscription task cancelled")
+            break
+        except Exception as e:
+            logger.warning(
+                "LLM registry subscription error: %s. Retrying in 5s...", e
+            )
+            await asyncio.sleep(5)

autogpt_platform/backend/backend/data/llm_registry/registry.py

@@ -0,0 +1,254 @@
+"""Core LLM registry implementation for managing models dynamically."""
+
+from __future__ import annotations
+
+import asyncio
+import logging
+from typing import Any
+
+import prisma.models
+from pydantic import BaseModel, ConfigDict
+
+from backend.blocks.llm import ModelMetadata
+from backend.util.cache import cached
+
+logger = logging.getLogger(__name__)
+
+
+class RegistryModelCost(BaseModel):
+    """Cost configuration for an LLM model."""
+
+    model_config = ConfigDict(frozen=True)
+
+    unit: str  # "RUN" or "TOKENS"
+    credit_cost: int
+    credential_provider: str
+    credential_id: str | None = None
+    credential_type: str | None = None
+    currency: str | None = None
+    metadata: dict[str, Any] = {}
+
+
+class RegistryModelCreator(BaseModel):
+    """Creator information for an LLM model."""
+
+    model_config = ConfigDict(frozen=True)
+
+    id: str
+    name: str
+    display_name: str
+    description: str | None = None
+    website_url: str | None = None
+    logo_url: str | None = None
+
+
+class RegistryModel(BaseModel):
+    """Represents a model in the LLM registry."""
+
+    model_config = ConfigDict(frozen=True)
+
+    slug: str
+    display_name: str
+    description: str | None = None
+    metadata: ModelMetadata
+    capabilities: dict[str, Any] = {}
+    extra_metadata: dict[str, Any] = {}
+    provider_display_name: str
+    is_enabled: bool
+    is_recommended: bool = False
+    costs: tuple[RegistryModelCost, ...] = ()
+    creator: RegistryModelCreator | None = None
+
+    # Typed capability fields from DB schema
+    supports_tools: bool = False
+    supports_json_output: bool = False
+    supports_reasoning: bool = False
+    supports_parallel_tool_calls: bool = False
+
+
+# L1 in-process cache — Redis is the shared L2 via @cached(shared_cache=True)
+_dynamic_models: dict[str, RegistryModel] = {}
+_schema_options: list[dict[str, str]] = []
+_lock = asyncio.Lock()
+
+
+def _record_to_registry_model(record: prisma.models.LlmModel) -> RegistryModel:  # type: ignore[name-defined]
+    """Transform a raw Prisma LlmModel record into a RegistryModel instance."""
+    costs = tuple(
+        RegistryModelCost(
+            unit=str(cost.unit),
+            credit_cost=cost.creditCost,
+            credential_provider=cost.credentialProvider,
+            credential_id=cost.credentialId,
+            credential_type=cost.credentialType,
+            currency=cost.currency,
+            metadata=dict(cost.metadata or {}),
+        )
+        for cost in (record.Costs or [])
+    )
+
+    creator = None
+    if record.Creator:
+        creator = RegistryModelCreator(
+            id=record.Creator.id,
+            name=record.Creator.name,
+            display_name=record.Creator.displayName,
+            description=record.Creator.description,
+            website_url=record.Creator.websiteUrl,
+            logo_url=record.Creator.logoUrl,
+        )
+
+    capabilities = dict(record.capabilities or {})
+
+    if not record.Provider:
+        logger.warning(
+            "LlmModel %s has no Provider despite NOT NULL FK - "
+            "falling back to providerId %s",
+            record.slug,
+            record.providerId,
+        )
+    provider_name = record.Provider.name if record.Provider else record.providerId
+    provider_display = (
+        record.Provider.displayName if record.Provider else record.providerId
+    )
+    creator_name = record.Creator.displayName if record.Creator else "Unknown"
+
+    if record.priceTier not in (1, 2, 3):
+        logger.warning(
+            "LlmModel %s has out-of-range priceTier=%s, defaulting to 1",
+            record.slug,
+            record.priceTier,
+        )
+    price_tier = record.priceTier if record.priceTier in (1, 2, 3) else 1
+
+    metadata = ModelMetadata(
+        provider=provider_name,
+        context_window=record.contextWindow,
+        max_output_tokens=(
+            record.maxOutputTokens
+            if record.maxOutputTokens is not None
+            else record.contextWindow
+        ),
+        display_name=record.displayName,
+        provider_name=provider_display,
+        creator_name=creator_name,
+        price_tier=price_tier,
+    )
+
+    return RegistryModel(
+        slug=record.slug,
+        display_name=record.displayName,
+        description=record.description,
+        metadata=metadata,
+        capabilities=capabilities,
+        extra_metadata=dict(record.metadata or {}),
+        provider_display_name=provider_display,
+        is_enabled=record.isEnabled,
+        is_recommended=record.isRecommended,
+        costs=costs,
+        creator=creator,
+        supports_tools=record.supportsTools,
+        supports_json_output=record.supportsJsonOutput,
+        supports_reasoning=record.supportsReasoning,
+        supports_parallel_tool_calls=record.supportsParallelToolCalls,
+    )
+
+
+@cached(maxsize=1, ttl_seconds=300, shared_cache=True, refresh_ttl_on_get=True)
+async def _fetch_registry_from_db() -> list[RegistryModel]:
+    """Fetch all LLM models from the database.
+
+    Results are cached in Redis (shared_cache=True) so subsequent calls within
+    the TTL window skip the DB entirely — both within this process and across
+    all other workers that share the same Redis instance.
+    """
+    records = await prisma.models.LlmModel.prisma().find_many(  # type: ignore[attr-defined]
+        include={"Provider": True, "Costs": True, "Creator": True}
+    )
+    logger.info("Fetched %d LLM models from database", len(records))
+    return [_record_to_registry_model(r) for r in records]
+
+
+def clear_registry_cache() -> None:
+    """Invalidate the shared Redis cache for the registry DB fetch.
+
+    Call this before refresh_llm_registry() after any admin DB mutation so the
+    next fetch hits the database rather than serving the now-stale cached data.
+    """
+    _fetch_registry_from_db.cache_clear()
+
+
+async def refresh_llm_registry() -> None:
+    """Refresh the in-process L1 cache from Redis/DB.
+
+    On the first call (or after clear_registry_cache()), fetches fresh data
+    from the database and stores it in Redis.  Subsequent calls by other
+    workers hit the Redis cache instead of the DB.
+    """
+    async with _lock:
+        try:
+            models = await _fetch_registry_from_db()
+            new_models = {m.slug: m for m in models}
+
+            global _dynamic_models, _schema_options
+            _dynamic_models = new_models
+            _schema_options = _build_schema_options()
+
+            logger.info(
+                "LLM registry refreshed: %d models, %d schema options",
+                len(_dynamic_models),
+                len(_schema_options),
+            )
+        except Exception as e:
+            logger.error("Failed to refresh LLM registry: %s", e, exc_info=True)
+            raise
+
+
+def _build_schema_options() -> list[dict[str, str]]:
+    """Build schema options for model selection dropdown. Only includes enabled models."""
+    return [
+        {
+            "label": model.display_name,
+            "value": model.slug,
+            "group": model.metadata.provider,
+            "description": model.description or "",
+        }
+        for model in sorted(
+            _dynamic_models.values(), key=lambda m: m.display_name.lower()
+        )
+        if model.is_enabled
+    ]
+
+
+def get_model(slug: str) -> RegistryModel | None:
+    """Get a model by slug from the registry."""
+    return _dynamic_models.get(slug)
+
+
+def get_all_models() -> list[RegistryModel]:
+    """Get all models from the registry (including disabled)."""
+    return list(_dynamic_models.values())
+
+
+def get_enabled_models() -> list[RegistryModel]:
+    """Get only enabled models from the registry."""
+    return [model for model in _dynamic_models.values() if model.is_enabled]
+
+
+def get_schema_options() -> list[dict[str, str]]:
+    """Get schema options for model selection dropdown (enabled models only)."""
+    return list(_schema_options)
+
+
+def get_default_model_slug() -> str | None:
+    """Get the default model slug (first recommended, or first enabled)."""
+    models = sorted(_dynamic_models.values(), key=lambda m: m.display_name)
+    recommended = next(
+        (m.slug for m in models if m.is_recommended and m.is_enabled), None
+    )
+    return recommended or next((m.slug for m in models if m.is_enabled), None)
+
+
+def get_all_model_slugs_for_validation() -> list[str]:
+    """Get all model slugs for validation (enabled models only)."""
+    return [model.slug for model in _dynamic_models.values() if model.is_enabled]

autogpt_platform/backend/backend/data/llm_registry/registry_test.py

@@ -0,0 +1,358 @@
+"""Unit tests for the LLM registry module."""
+
+from __future__ import annotations
+
+import asyncio
+from unittest.mock import AsyncMock, Mock, patch
+
+import pytest
+import pydantic
+
+from backend.data.llm_registry.registry import (
+    RegistryModel,
+    RegistryModelCost,
+    RegistryModelCreator,
+    _build_schema_options,
+    _record_to_registry_model,
+    get_default_model_slug,
+    get_schema_options,
+    refresh_llm_registry,
+)
+
+
+# ---------------------------------------------------------------------------
+# Helpers
+# ---------------------------------------------------------------------------
+
+
+def _make_mock_record(**overrides):
+    """Build a realistic mock Prisma LlmModel record."""
+    provider = Mock()
+    provider.name = "openai"
+    provider.displayName = "OpenAI"
+
+    record = Mock()
+    record.slug = "openai/gpt-4o"
+    record.displayName = "GPT-4o"
+    record.description = "Latest GPT model"
+    record.providerId = "provider-uuid"
+    record.Provider = provider
+    record.creatorId = "creator-uuid"
+    record.Creator = None
+    record.contextWindow = 128000
+    record.maxOutputTokens = 16384
+    record.priceTier = 2
+    record.isEnabled = True
+    record.isRecommended = False
+    record.supportsTools = True
+    record.supportsJsonOutput = True
+    record.supportsReasoning = False
+    record.supportsParallelToolCalls = True
+    record.capabilities = {}
+    record.metadata = {}
+    record.Costs = []
+
+    for key, value in overrides.items():
+        setattr(record, key, value)
+    return record
+
+
+def _make_registry_model(**kwargs) -> RegistryModel:
+    """Build a minimal RegistryModel for testing registry-level functions."""
+    from backend.blocks.llm import ModelMetadata
+
+    defaults = dict(
+        slug="openai/gpt-4o",
+        display_name="GPT-4o",
+        description=None,
+        metadata=ModelMetadata(
+            provider="openai",
+            context_window=128000,
+            max_output_tokens=16384,
+            display_name="GPT-4o",
+            provider_name="OpenAI",
+            creator_name="Unknown",
+            price_tier=2,
+        ),
+        capabilities={},
+        extra_metadata={},
+        provider_display_name="OpenAI",
+        is_enabled=True,
+        is_recommended=False,
+    )
+    defaults.update(kwargs)
+    return RegistryModel(**defaults)
+
+
+# ---------------------------------------------------------------------------
+# _record_to_registry_model tests
+# ---------------------------------------------------------------------------
+
+
+def test_record_to_registry_model():
+    """Happy-path: well-formed record produces a correct RegistryModel."""
+    record = _make_mock_record()
+    model = _record_to_registry_model(record)
+
+    assert model.slug == "openai/gpt-4o"
+    assert model.display_name == "GPT-4o"
+    assert model.description == "Latest GPT model"
+    assert model.provider_display_name == "OpenAI"
+    assert model.is_enabled is True
+    assert model.is_recommended is False
+    assert model.supports_tools is True
+    assert model.supports_json_output is True
+    assert model.supports_reasoning is False
+    assert model.supports_parallel_tool_calls is True
+    assert model.metadata.provider == "openai"
+    assert model.metadata.context_window == 128000
+    assert model.metadata.max_output_tokens == 16384
+    assert model.metadata.price_tier == 2
+    assert model.creator is None
+    assert model.costs == ()
+
+
+def test_record_to_registry_model_missing_provider(caplog):
+    """Record with no Provider relation falls back to providerId and logs a warning."""
+    record = _make_mock_record(Provider=None, providerId="provider-uuid")
+    with caplog.at_level("WARNING"):
+        model = _record_to_registry_model(record)
+
+    assert "no Provider" in caplog.text
+    assert model.metadata.provider == "provider-uuid"
+    assert model.provider_display_name == "provider-uuid"
+
+
+def test_record_to_registry_model_missing_creator():
+    """When Creator is None, creator_name defaults to 'Unknown' and creator field is None."""
+    record = _make_mock_record(Creator=None)
+    model = _record_to_registry_model(record)
+
+    assert model.creator is None
+    assert model.metadata.creator_name == "Unknown"
+
+
+def test_record_to_registry_model_with_creator():
+    """When Creator is present, it is parsed into RegistryModelCreator."""
+    creator_mock = Mock()
+    creator_mock.id = "creator-uuid"
+    creator_mock.name = "openai"
+    creator_mock.displayName = "OpenAI"
+    creator_mock.description = "AI company"
+    creator_mock.websiteUrl = "https://openai.com"
+    creator_mock.logoUrl = "https://openai.com/logo.png"
+
+    record = _make_mock_record(Creator=creator_mock)
+    model = _record_to_registry_model(record)
+
+    assert model.creator is not None
+    assert isinstance(model.creator, RegistryModelCreator)
+    assert model.creator.id == "creator-uuid"
+    assert model.creator.display_name == "OpenAI"
+    assert model.metadata.creator_name == "OpenAI"
+
+
+def test_record_to_registry_model_null_max_output_tokens():
+    """maxOutputTokens=None falls back to contextWindow."""
+    record = _make_mock_record(maxOutputTokens=None, contextWindow=64000)
+    model = _record_to_registry_model(record)
+
+    assert model.metadata.max_output_tokens == 64000
+
+
+def test_record_to_registry_model_invalid_price_tier(caplog):
+    """Out-of-range priceTier is coerced to 1 and a warning is logged."""
+    record = _make_mock_record(priceTier=99)
+    with caplog.at_level("WARNING"):
+        model = _record_to_registry_model(record)
+
+    assert "out-of-range priceTier" in caplog.text
+    assert model.metadata.price_tier == 1
+
+
+def test_record_to_registry_model_with_costs():
+    """Costs are parsed into RegistryModelCost tuples."""
+    cost_mock = Mock()
+    cost_mock.unit = "TOKENS"
+    cost_mock.creditCost = 10
+    cost_mock.credentialProvider = "openai"
+    cost_mock.credentialId = None
+    cost_mock.credentialType = None
+    cost_mock.currency = "USD"
+    cost_mock.metadata = {}
+
+    record = _make_mock_record(Costs=[cost_mock])
+    model = _record_to_registry_model(record)
+
+    assert len(model.costs) == 1
+    cost = model.costs[0]
+    assert isinstance(cost, RegistryModelCost)
+    assert cost.unit == "TOKENS"
+    assert cost.credit_cost == 10
+    assert cost.credential_provider == "openai"
+
+
+# ---------------------------------------------------------------------------
+# get_default_model_slug tests
+# ---------------------------------------------------------------------------
+
+
+def test_get_default_model_slug_recommended():
+    """Recommended model is preferred over non-recommended enabled models."""
+    import backend.data.llm_registry.registry as reg
+
+    reg._dynamic_models = {
+        "openai/gpt-4o": _make_registry_model(
+            slug="openai/gpt-4o", display_name="GPT-4o", is_recommended=False
+        ),
+        "openai/gpt-4o-recommended": _make_registry_model(
+            slug="openai/gpt-4o-recommended",
+            display_name="GPT-4o Recommended",
+            is_recommended=True,
+        ),
+    }
+
+    result = get_default_model_slug()
+    assert result == "openai/gpt-4o-recommended"
+
+
+def test_get_default_model_slug_fallback():
+    """With no recommended model, falls back to first enabled (alphabetical)."""
+    import backend.data.llm_registry.registry as reg
+
+    reg._dynamic_models = {
+        "openai/gpt-4o": _make_registry_model(
+            slug="openai/gpt-4o", display_name="GPT-4o", is_recommended=False
+        ),
+        "openai/gpt-3.5": _make_registry_model(
+            slug="openai/gpt-3.5", display_name="GPT-3.5", is_recommended=False
+        ),
+    }
+
+    result = get_default_model_slug()
+    # Sorted alphabetically: GPT-3.5 < GPT-4o
+    assert result == "openai/gpt-3.5"
+
+
+def test_get_default_model_slug_empty():
+    """Empty registry returns None."""
+    import backend.data.llm_registry.registry as reg
+
+    reg._dynamic_models = {}
+
+    result = get_default_model_slug()
+    assert result is None
+
+
+# ---------------------------------------------------------------------------
+# _build_schema_options / get_schema_options tests
+# ---------------------------------------------------------------------------
+
+
+def test_build_schema_options():
+    """Only enabled models appear, sorted case-insensitively."""
+    import backend.data.llm_registry.registry as reg
+
+    reg._dynamic_models = {
+        "openai/gpt-4o": _make_registry_model(
+            slug="openai/gpt-4o", display_name="GPT-4o", is_enabled=True
+        ),
+        "openai/disabled": _make_registry_model(
+            slug="openai/disabled", display_name="Disabled Model", is_enabled=False
+        ),
+        "openai/gpt-3.5": _make_registry_model(
+            slug="openai/gpt-3.5", display_name="gpt-3.5", is_enabled=True
+        ),
+    }
+
+    options = _build_schema_options()
+    slugs = [o["value"] for o in options]
+
+    # disabled model should be excluded
+    assert "openai/disabled" not in slugs
+    # only enabled models
+    assert "openai/gpt-4o" in slugs
+    assert "openai/gpt-3.5" in slugs
+    # case-insensitive sort: "gpt-3.5" < "GPT-4o" (both lowercase: "gpt-3.5" < "gpt-4o")
+    assert slugs.index("openai/gpt-3.5") < slugs.index("openai/gpt-4o")
+
+    # Verify structure
+    for option in options:
+        assert "label" in option
+        assert "value" in option
+        assert "group" in option
+        assert "description" in option
+
+
+def test_get_schema_options_returns_copy():
+    """Mutating the returned list does not affect the internal cache."""
+    import backend.data.llm_registry.registry as reg
+
+    reg._dynamic_models = {
+        "openai/gpt-4o": _make_registry_model(slug="openai/gpt-4o", display_name="GPT-4o"),
+    }
+    reg._schema_options = _build_schema_options()
+
+    options = get_schema_options()
+    original_length = len(options)
+    options.append({"label": "Injected", "value": "evil/model", "group": "evil", "description": ""})
+
+    # Internal state should be unchanged
+    assert len(get_schema_options()) == original_length
+
+
+# ---------------------------------------------------------------------------
+# Pydantic frozen model tests
+# ---------------------------------------------------------------------------
+
+
+def test_registry_model_frozen():
+    """Pydantic frozen=True should reject attribute assignment."""
+    model = _make_registry_model()
+
+    with pytest.raises((pydantic.ValidationError, TypeError)):
+        model.slug = "changed/slug"  # type: ignore[misc]
+
+
+def test_registry_model_cost_frozen():
+    """RegistryModelCost is also frozen."""
+    cost = RegistryModelCost(
+        unit="TOKENS",
+        credit_cost=5,
+        credential_provider="openai",
+    )
+    with pytest.raises((pydantic.ValidationError, TypeError)):
+        cost.unit = "RUN"  # type: ignore[misc]
+
+
+# ---------------------------------------------------------------------------
+# refresh_llm_registry tests
+# ---------------------------------------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_refresh_llm_registry():
+    """Mock prisma find_many, verify cache is populated after refresh."""
+    import backend.data.llm_registry.registry as reg
+
+    record = _make_mock_record()
+    mock_find_many = AsyncMock(return_value=[record])
+
+    with patch("prisma.models.LlmModel.prisma") as mock_prisma_cls:
+        mock_prisma_instance = Mock()
+        mock_prisma_instance.find_many = mock_find_many
+        mock_prisma_cls.return_value = mock_prisma_instance
+
+        # Clear state first
+        reg._dynamic_models = {}
+        reg._schema_options = []
+
+        await refresh_llm_registry()
+
+    assert "openai/gpt-4o" in reg._dynamic_models
+    model = reg._dynamic_models["openai/gpt-4o"]
+    assert isinstance(model, RegistryModel)
+    assert model.slug == "openai/gpt-4o"
+    # Schema options should be populated too
+    assert len(reg._schema_options) == 1
+    assert reg._schema_options[0]["value"] == "openai/gpt-4o"

autogpt_platform/backend/backend/server/v2/llm/__init__.py

@@ -0,0 +1,5 @@
+"""LLM registry public API."""
+
+from .routes import router
+
+__all__ = ["router"]

autogpt_platform/backend/backend/server/v2/llm/model.py

@@ -0,0 +1,68 @@
+"""Pydantic models for LLM registry public API."""
+
+from __future__ import annotations
+
+from typing import Any
+
+import pydantic
+
+
+class LlmModelCost(pydantic.BaseModel):
+    """Cost configuration for an LLM model."""
+
+    unit: str  # "RUN" or "TOKENS"
+    credit_cost: int = pydantic.Field(ge=0)
+    credential_provider: str
+    credential_id: str | None = None
+    credential_type: str | None = None
+    currency: str | None = None
+    metadata: dict[str, Any] = pydantic.Field(default_factory=dict)
+
+
+class LlmModelCreator(pydantic.BaseModel):
+    """Represents the organization that created/trained the model."""
+
+    id: str
+    name: str
+    display_name: str
+    description: str | None = None
+    website_url: str | None = None
+    logo_url: str | None = None
+
+
+class LlmModel(pydantic.BaseModel):
+    """Public-facing LLM model information."""
+
+    slug: str
+    display_name: str
+    description: str | None = None
+    provider_name: str
+    creator: LlmModelCreator | None = None
+    context_window: int
+    max_output_tokens: int | None = None
+    price_tier: int  # 1=cheapest, 2=medium, 3=expensive
+    is_enabled: bool = True
+    is_recommended: bool = False
+    capabilities: dict[str, Any] = pydantic.Field(default_factory=dict)
+    costs: list[LlmModelCost] = pydantic.Field(default_factory=list)
+
+
+class LlmProvider(pydantic.BaseModel):
+    """Provider with its enabled models."""
+
+    name: str
+    display_name: str
+    models: list[LlmModel] = pydantic.Field(default_factory=list)
+
+
+class LlmModelsResponse(pydantic.BaseModel):
+    """Response for GET /llm/models."""
+
+    models: list[LlmModel]
+    total: int
+
+
+class LlmProvidersResponse(pydantic.BaseModel):
+    """Response for GET /llm/providers."""
+
+    providers: list[LlmProvider]

autogpt_platform/backend/backend/server/v2/llm/routes.py

@@ -0,0 +1,143 @@
+"""Public read-only API for LLM registry."""
+
+import autogpt_libs.auth
+import fastapi
+
+from backend.data.llm_registry import (
+    RegistryModelCreator,
+    get_all_models,
+    get_enabled_models,
+)
+from backend.server.v2.llm import model as llm_model
+
+router = fastapi.APIRouter(
+    prefix="/llm",
+    tags=["llm"],
+    dependencies=[fastapi.Security(autogpt_libs.auth.requires_user)],
+)
+
+
+def _map_creator(
+    creator: RegistryModelCreator | None,
+) -> llm_model.LlmModelCreator | None:
+    """Convert registry creator to API model."""
+    if not creator:
+        return None
+    return llm_model.LlmModelCreator(
+        id=creator.id,
+        name=creator.name,
+        display_name=creator.display_name,
+        description=creator.description,
+        website_url=creator.website_url,
+        logo_url=creator.logo_url,
+    )
+
+
+@router.get("/models", response_model=llm_model.LlmModelsResponse)
+async def list_models(
+    enabled_only: bool = fastapi.Query(
+        default=True, description="Only return enabled models"
+    ),
+):
+    """
+    List all LLM models available to users.
+
+    Returns models from the in-memory registry cache.
+    Use enabled_only=true to filter to only enabled models (default).
+    """
+    # Get models from in-memory registry
+    registry_models = get_enabled_models() if enabled_only else get_all_models()
+
+    # Map to API response models
+    models = [
+        llm_model.LlmModel(
+            slug=model.slug,
+            display_name=model.display_name,
+            description=model.description,
+            provider_name=model.provider_display_name,
+            creator=_map_creator(model.creator),
+            context_window=model.metadata.context_window,
+            max_output_tokens=model.metadata.max_output_tokens,
+            price_tier=model.metadata.price_tier,
+            is_enabled=model.is_enabled,
+            is_recommended=model.is_recommended,
+            capabilities=model.capabilities,
+            costs=[
+                llm_model.LlmModelCost(
+                    unit=cost.unit,
+                    credit_cost=cost.credit_cost,
+                    credential_provider=cost.credential_provider,
+                    credential_id=cost.credential_id,
+                    credential_type=cost.credential_type,
+                    currency=cost.currency,
+                    metadata=cost.metadata,
+                )
+                for cost in model.costs
+            ],
+        )
+        for model in registry_models
+    ]
+
+    return llm_model.LlmModelsResponse(models=models, total=len(models))
+
+
+@router.get("/providers", response_model=llm_model.LlmProvidersResponse)
+async def list_providers():
+    """
+    List all LLM providers with their enabled models.
+
+    Groups enabled models by provider from the in-memory registry.
+    """
+    # Get all enabled models and group by provider
+    registry_models = get_enabled_models()
+
+    # Group models by provider
+    provider_map: dict[str, list] = {}
+    for model in registry_models:
+        provider_key = model.metadata.provider
+        if provider_key not in provider_map:
+            provider_map[provider_key] = []
+        provider_map[provider_key].append(model)
+
+    # Build provider responses
+    providers = []
+    for provider_key, models in sorted(provider_map.items()):
+        # Use the first model's provider display name
+        display_name = models[0].provider_display_name if models else provider_key
+
+        providers.append(
+            llm_model.LlmProvider(
+                name=provider_key,
+                display_name=display_name,
+                models=[
+                    llm_model.LlmModel(
+                        slug=model.slug,
+                        display_name=model.display_name,
+                        description=model.description,
+                        provider_name=model.provider_display_name,
+                        creator=_map_creator(model.creator),
+                        context_window=model.metadata.context_window,
+                        max_output_tokens=model.metadata.max_output_tokens,
+                        price_tier=model.metadata.price_tier,
+                        is_enabled=model.is_enabled,
+            is_recommended=model.is_recommended,
+                        capabilities=model.capabilities,
+                        costs=[
+                            llm_model.LlmModelCost(
+                                unit=cost.unit,
+                                credit_cost=cost.credit_cost,
+                                credential_provider=cost.credential_provider,
+                                credential_id=cost.credential_id,
+                                credential_type=cost.credential_type,
+                                currency=cost.currency,
+                                metadata=cost.metadata,
+                            )
+                            for cost in model.costs
+                        ],
+                    )
+                    for model in sorted(models, key=lambda m: m.display_name)
+                ],
+            )
+        )
+
+    return llm_model.LlmProvidersResponse(providers=providers)

autogpt_platform/backend/migrations/20260310_add_llm_registry_schema/migration.sql

@@ -0,0 +1,148 @@
+-- CreateEnum
+CREATE TYPE "LlmCostUnit" AS ENUM ('RUN', 'TOKENS');
+
+-- CreateTable
+CREATE TABLE "LlmProvider" (
+    "id" TEXT NOT NULL,
+    "createdAt" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    "updatedAt" TIMESTAMP(3) NOT NULL,
+    "name" TEXT NOT NULL,
+    "displayName" TEXT NOT NULL,
+    "description" TEXT,
+    "defaultCredentialProvider" TEXT,
+    "defaultCredentialId" TEXT,
+    "defaultCredentialType" TEXT,
+    "metadata" JSONB NOT NULL DEFAULT '{}',
+
+    CONSTRAINT "LlmProvider_pkey" PRIMARY KEY ("id")
+);
+
+-- CreateTable
+CREATE TABLE "LlmModelCreator" (
+    "id" TEXT NOT NULL,
+    "createdAt" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    "updatedAt" TIMESTAMP(3) NOT NULL,
+    "name" TEXT NOT NULL,
+    "displayName" TEXT NOT NULL,
+    "description" TEXT,
+    "websiteUrl" TEXT,
+    "logoUrl" TEXT,
+    "metadata" JSONB NOT NULL DEFAULT '{}',
+
+    CONSTRAINT "LlmModelCreator_pkey" PRIMARY KEY ("id")
+);
+
+-- CreateTable
+CREATE TABLE "LlmModel" (
+    "id" TEXT NOT NULL,
+    "createdAt" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    "updatedAt" TIMESTAMP(3) NOT NULL,
+    "slug" TEXT NOT NULL,
+    "displayName" TEXT NOT NULL,
+    "description" TEXT,
+    "providerId" TEXT NOT NULL,
+    "creatorId" TEXT,
+    "contextWindow" INTEGER NOT NULL,
+    "maxOutputTokens" INTEGER,
+    "priceTier" INTEGER NOT NULL DEFAULT 1,
+    "isEnabled" BOOLEAN NOT NULL DEFAULT true,
+    "isRecommended" BOOLEAN NOT NULL DEFAULT false,
+    "supportsTools" BOOLEAN NOT NULL DEFAULT false,
+    "supportsJsonOutput" BOOLEAN NOT NULL DEFAULT false,
+    "supportsReasoning" BOOLEAN NOT NULL DEFAULT false,
+    "supportsParallelToolCalls" BOOLEAN NOT NULL DEFAULT false,
+    "capabilities" JSONB NOT NULL DEFAULT '{}',
+    "metadata" JSONB NOT NULL DEFAULT '{}',
+
+    CONSTRAINT "LlmModel_pkey" PRIMARY KEY ("id")
+);
+
+-- CreateTable
+CREATE TABLE "LlmModelCost" (
+    "id" TEXT NOT NULL,
+    "createdAt" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    "updatedAt" TIMESTAMP(3) NOT NULL,
+    "unit" "LlmCostUnit" NOT NULL DEFAULT 'RUN',
+    "creditCost" INTEGER NOT NULL,
+    "credentialProvider" TEXT NOT NULL,
+    "credentialId" TEXT,
+    "credentialType" TEXT,
+    "currency" TEXT,
+    "metadata" JSONB NOT NULL DEFAULT '{}',
+    "llmModelId" TEXT NOT NULL,
+
+    CONSTRAINT "LlmModelCost_pkey" PRIMARY KEY ("id")
+);
+
+-- CreateTable
+CREATE TABLE "LlmModelMigration" (
+    "id" TEXT NOT NULL,
+    "createdAt" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    "updatedAt" TIMESTAMP(3) NOT NULL,
+    "sourceModelSlug" TEXT NOT NULL,
+    "targetModelSlug" TEXT NOT NULL,
+    "reason" TEXT,
+    "migratedNodeIds" JSONB NOT NULL DEFAULT '[]',
+    "nodeCount" INTEGER NOT NULL,
+    "customCreditCost" INTEGER,
+    "isReverted" BOOLEAN NOT NULL DEFAULT false,
+    "revertedAt" TIMESTAMP(3),
+
+    CONSTRAINT "LlmModelMigration_pkey" PRIMARY KEY ("id")
+);
+
+-- CreateIndex
+CREATE UNIQUE INDEX "LlmProvider_name_key" ON "LlmProvider"("name");
+
+-- CreateIndex
+CREATE UNIQUE INDEX "LlmModelCreator_name_key" ON "LlmModelCreator"("name");
+
+-- CreateIndex
+CREATE UNIQUE INDEX "LlmModel_slug_key" ON "LlmModel"("slug");
+
+-- CreateIndex
+CREATE INDEX "LlmModel_providerId_isEnabled_idx" ON "LlmModel"("providerId", "isEnabled");
+
+-- CreateIndex
+CREATE INDEX "LlmModel_creatorId_idx" ON "LlmModel"("creatorId");
+
+-- CreateIndex (partial unique for default costs - no specific credential)
+CREATE UNIQUE INDEX "LlmModelCost_default_cost_key" ON "LlmModelCost"("llmModelId", "credentialProvider", "unit") WHERE "credentialId" IS NULL;
+
+-- CreateIndex (partial unique for credential-specific costs)
+CREATE UNIQUE INDEX "LlmModelCost_credential_cost_key" ON "LlmModelCost"("llmModelId", "credentialProvider", "credentialId", "unit") WHERE "credentialId" IS NOT NULL;
+
+-- CreateIndex
+CREATE INDEX "LlmModelMigration_targetModelSlug_idx" ON "LlmModelMigration"("targetModelSlug");
+
+-- CreateIndex
+CREATE INDEX "LlmModelMigration_sourceModelSlug_isReverted_idx" ON "LlmModelMigration"("sourceModelSlug", "isReverted");
+
+-- CreateIndex (partial unique to prevent multiple active migrations per source)
+CREATE UNIQUE INDEX "LlmModelMigration_active_source_key" ON "LlmModelMigration"("sourceModelSlug") WHERE "isReverted" = false;
+
+-- AddForeignKey
+ALTER TABLE "LlmModel" ADD CONSTRAINT "LlmModel_providerId_fkey" FOREIGN KEY ("providerId") REFERENCES "LlmProvider"("id") ON DELETE RESTRICT ON UPDATE CASCADE;
+
+-- AddForeignKey
+ALTER TABLE "LlmModel" ADD CONSTRAINT "LlmModel_creatorId_fkey" FOREIGN KEY ("creatorId") REFERENCES "LlmModelCreator"("id") ON DELETE SET NULL ON UPDATE CASCADE;
+
+-- AddForeignKey
+ALTER TABLE "LlmModelCost" ADD CONSTRAINT "LlmModelCost_llmModelId_fkey" FOREIGN KEY ("llmModelId") REFERENCES "LlmModel"("id") ON DELETE CASCADE ON UPDATE CASCADE;
+
+-- AddForeignKey
+ALTER TABLE "LlmModelMigration" ADD CONSTRAINT "LlmModelMigration_sourceModelSlug_fkey" FOREIGN KEY ("sourceModelSlug") REFERENCES "LlmModel"("slug") ON DELETE RESTRICT ON UPDATE CASCADE;
+
+-- AddForeignKey
+ALTER TABLE "LlmModelMigration" ADD CONSTRAINT "LlmModelMigration_targetModelSlug_fkey" FOREIGN KEY ("targetModelSlug") REFERENCES "LlmModel"("slug") ON DELETE RESTRICT ON UPDATE CASCADE;
+
+-- AddCheckConstraints (enforce data integrity)
+ALTER TABLE "LlmModel"
+    ADD CONSTRAINT "LlmModel_priceTier_check" CHECK ("priceTier" BETWEEN 1 AND 3);
+
+ALTER TABLE "LlmModelCost"
+    ADD CONSTRAINT "LlmModelCost_creditCost_check" CHECK ("creditCost" >= 0);
+
+ALTER TABLE "LlmModelMigration"
+    ADD CONSTRAINT "LlmModelMigration_nodeCount_check" CHECK ("nodeCount" >= 0),
+    ADD CONSTRAINT "LlmModelMigration_customCreditCost_check" CHECK ("customCreditCost" IS NULL OR "customCreditCost" >= 0);

autogpt_platform/backend/migrations/20260310_seed_llm_registry/migration.sql

@@ -0,0 +1,287 @@
+-- Seed LLM Registry from existing hard-coded data
+-- This migration populates the LlmProvider, LlmModelCreator, LlmModel, and LlmModelCost tables
+-- with data from the existing MODEL_METADATA and MODEL_COST dictionaries
+
+-- Insert Providers
+INSERT INTO "LlmProvider" ("id", "createdAt", "updatedAt", "name", "displayName", "description", "defaultCredentialProvider", "defaultCredentialType", "metadata")
+VALUES
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'openai', 'OpenAI', 'OpenAI language models', 'openai', 'api_key', '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'anthropic', 'Anthropic', 'Anthropic Claude models', 'anthropic', 'api_key', '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'groq', 'Groq', 'Groq inference API', 'groq', 'api_key', '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'open_router', 'OpenRouter', 'OpenRouter unified API', 'open_router', 'api_key', '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'aiml_api', 'AI/ML API', 'AI/ML API models', 'aiml_api', 'api_key', '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'ollama', 'Ollama', 'Ollama local models', 'ollama', 'api_key', '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'llama_api', 'Llama API', 'Llama API models', 'llama_api', 'api_key', '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'v0', 'v0', 'v0 by Vercel models', 'v0', 'api_key', '{}'::jsonb)
+ON CONFLICT ("name") DO NOTHING;
+
+-- Insert Model Creators
+INSERT INTO "LlmModelCreator" ("id", "createdAt", "updatedAt", "name", "displayName", "description", "websiteUrl", "logoUrl", "metadata")
+VALUES
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'openai', 'OpenAI', 'Creator of GPT, O1, O3, and DALL-E models', 'https://openai.com', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'anthropic', 'Anthropic', 'Creator of Claude AI models', 'https://anthropic.com', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'meta', 'Meta', 'Creator of Llama foundation models', 'https://llama.meta.com', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'google', 'Google', 'Creator of Gemini and PaLM models', 'https://deepmind.google', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'mistralai', 'Mistral AI', 'Creator of Mistral and Codestral models', 'https://mistral.ai', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'cohere', 'Cohere', 'Creator of Command language models', 'https://cohere.com', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'deepseek', 'DeepSeek', 'Creator of DeepSeek reasoning models', 'https://deepseek.com', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'alibaba', 'Alibaba', 'Creator of Qwen language models', 'https://qwenlm.github.io', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'nvidia', 'NVIDIA', 'Creator of Nemotron models', 'https://nvidia.com', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'vercel', 'Vercel', 'Creator of v0 AI models', 'https://v0.dev', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'microsoft', 'Microsoft', 'Creator of Phi models', 'https://microsoft.com', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'xai', 'xAI', 'Creator of Grok models', 'https://x.ai', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'perplexity', 'Perplexity AI', 'Creator of Sonar search models', 'https://perplexity.ai', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'nousresearch', 'Nous Research', 'Creator of Hermes language models', 'https://nousresearch.com', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'amazon', 'Amazon', 'Creator of Nova language models', 'https://aws.amazon.com', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'gryphe', 'Gryphe', 'Creator of MythoMax models', 'https://huggingface.co/Gryphe', NULL, '{}'::jsonb),
+    (gen_random_uuid(), CURRENT_TIMESTAMP, CURRENT_TIMESTAMP, 'moonshotai', 'Moonshot AI', 'Creator of Kimi language models', 'https://moonshot.ai', NULL, '{}'::jsonb)
+ON CONFLICT ("name") DO NOTHING;
+
+-- Insert Models (using CTEs to reference provider and creator IDs)
+WITH provider_ids AS (
+    SELECT "id", "name" FROM "LlmProvider"
+),
+creator_ids AS (
+    SELECT "id", "name" FROM "LlmModelCreator"
+)
+INSERT INTO "LlmModel" ("id", "createdAt", "updatedAt", "slug", "displayName", "description", "providerId", "creatorId", "contextWindow", "maxOutputTokens", "isEnabled", "capabilities", "metadata")
+SELECT
+    gen_random_uuid(),
+    CURRENT_TIMESTAMP,
+    CURRENT_TIMESTAMP,
+    model_slug,
+    model_display_name,
+    NULL,
+    p."id",
+    c."id",
+    context_window,
+    max_output_tokens,
+    true,
+    '{}'::jsonb,
+    '{}'::jsonb
+FROM (VALUES
+    -- OpenAI models (creator: openai)
+    ('o3-2025-04-16', 'O3', 'openai', 'openai', 200000, 100000),
+    ('o3-mini', 'O3 Mini', 'openai', 'openai', 200000, 100000),
+    ('o1', 'O1', 'openai', 'openai', 200000, 100000),
+    ('o1-mini', 'O1 Mini', 'openai', 'openai', 128000, 65536),
+    ('gpt-5.2-2025-12-11', 'GPT-5.2', 'openai', 'openai', 400000, 128000),
+    ('gpt-5-2025-08-07', 'GPT 5', 'openai', 'openai', 400000, 128000),
+    ('gpt-5.1-2025-11-13', 'GPT 5.1', 'openai', 'openai', 400000, 128000),
+    ('gpt-5-mini-2025-08-07', 'GPT 5 Mini', 'openai', 'openai', 400000, 128000),
+    ('gpt-5-nano-2025-08-07', 'GPT 5 Nano', 'openai', 'openai', 400000, 128000),
+    ('gpt-5-chat-latest', 'GPT 5 Chat', 'openai', 'openai', 400000, 16384),
+    ('gpt-4.1-2025-04-14', 'GPT 4.1', 'openai', 'openai', 1000000, 32768),
+    ('gpt-4.1-mini-2025-04-14', 'GPT 4.1 Mini', 'openai', 'openai', 1047576, 32768),
+    ('gpt-4o-mini', 'GPT 4o Mini', 'openai', 'openai', 128000, 16384),
+    ('gpt-4o', 'GPT 4o', 'openai', 'openai', 128000, 16384),
+    ('gpt-4-turbo', 'GPT 4 Turbo', 'openai', 'openai', 128000, 4096),
+    -- Anthropic models (creator: anthropic)
+    ('claude-opus-4-6', 'Claude Opus 4.6', 'anthropic', 'anthropic', 200000, 128000),
+    ('claude-sonnet-4-6', 'Claude Sonnet 4.6', 'anthropic', 'anthropic', 200000, 64000),
+    ('claude-opus-4-1-20250805', 'Claude 4.1 Opus', 'anthropic', 'anthropic', 200000, 32000),
+    ('claude-opus-4-20250514', 'Claude 4 Opus', 'anthropic', 'anthropic', 200000, 32000),
+    ('claude-sonnet-4-20250514', 'Claude 4 Sonnet', 'anthropic', 'anthropic', 200000, 64000),
+    ('claude-opus-4-5-20251101', 'Claude 4.5 Opus', 'anthropic', 'anthropic', 200000, 64000),
+    ('claude-sonnet-4-5-20250929', 'Claude 4.5 Sonnet', 'anthropic', 'anthropic', 200000, 64000),
+    ('claude-haiku-4-5-20251001', 'Claude 4.5 Haiku', 'anthropic', 'anthropic', 200000, 64000),
+    ('claude-3-haiku-20240307', 'Claude 3 Haiku', 'anthropic', 'anthropic', 200000, 4096),
+    -- AI/ML API models (creators: alibaba, nvidia, meta)
+    ('Qwen/Qwen2.5-72B-Instruct-Turbo', 'Qwen 2.5 72B', 'aiml_api', 'alibaba', 32000, 8000),
+    ('nvidia/llama-3.1-nemotron-70b-instruct', 'Llama 3.1 Nemotron 70B', 'aiml_api', 'nvidia', 128000, 40000),
+    ('meta-llama/Llama-3.3-70B-Instruct-Turbo', 'Llama 3.3 70B', 'aiml_api', 'meta', 128000, NULL),
+    ('meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo', 'Meta Llama 3.1 70B', 'aiml_api', 'meta', 131000, 2000),
+    ('meta-llama/Llama-3.2-3B-Instruct-Turbo', 'Llama 3.2 3B', 'aiml_api', 'meta', 128000, NULL),
+    -- Groq models (creator: meta for Llama)
+    ('llama-3.3-70b-versatile', 'Llama 3.3 70B', 'groq', 'meta', 128000, 32768),
+    ('llama-3.1-8b-instant', 'Llama 3.1 8B', 'groq', 'meta', 128000, 8192),
+    -- Ollama models (creators: meta for Llama, mistralai for Mistral)
+    ('llama3.3', 'Llama 3.3', 'ollama', 'meta', 8192, NULL),
+    ('llama3.2', 'Llama 3.2', 'ollama', 'meta', 8192, NULL),
+    ('llama3', 'Llama 3', 'ollama', 'meta', 8192, NULL),
+    ('llama3.1:405b', 'Llama 3.1 405B', 'ollama', 'meta', 8192, NULL),
+    ('dolphin-mistral:latest', 'Dolphin Mistral', 'ollama', 'mistralai', 32768, NULL),
+    -- OpenRouter models (creators: google, mistralai, cohere, deepseek, perplexity, nousresearch, openai, amazon, microsoft, gryphe, meta, xai, moonshotai, alibaba)
+    ('google/gemini-2.5-pro-preview-03-25', 'Gemini 2.5 Pro', 'open_router', 'google', 1050000, 8192),
+    ('google/gemini-2.5-pro', 'Gemini 2.5 Pro', 'open_router', 'google', 1048576, 65536),
+    ('google/gemini-3.1-pro-preview', 'Gemini 3.1 Pro Preview', 'open_router', 'google', 1048576, 65536),
+    ('google/gemini-3-flash-preview', 'Gemini 3 Flash Preview', 'open_router', 'google', 1048576, 65536),
+    ('google/gemini-2.5-flash', 'Gemini 2.5 Flash', 'open_router', 'google', 1048576, 65535),
+    ('google/gemini-2.0-flash-001', 'Gemini 2.0 Flash', 'open_router', 'google', 1048576, 8192),
+    ('google/gemini-3.1-flash-lite-preview', 'Gemini 3.1 Flash Lite Preview', 'open_router', 'google', 1048576, 65536),
+    ('google/gemini-2.5-flash-lite-preview-06-17', 'Gemini 2.5 Flash Lite Preview', 'open_router', 'google', 1048576, 65535),
+    ('google/gemini-2.0-flash-lite-001', 'Gemini 2.0 Flash Lite', 'open_router', 'google', 1048576, 8192),
+    ('mistralai/mistral-nemo', 'Mistral Nemo', 'open_router', 'mistralai', 128000, 4096),
+    ('mistralai/mistral-large-2512', 'Mistral Large 3 2512', 'open_router', 'mistralai', 262144, NULL),
+    ('mistralai/mistral-medium-3.1', 'Mistral Medium 3.1', 'open_router', 'mistralai', 131072, NULL),
+    ('mistralai/mistral-small-3.2-24b-instruct', 'Mistral Small 3.2 24B', 'open_router', 'mistralai', 131072, 131072),
+    ('mistralai/codestral-2508', 'Codestral 2508', 'open_router', 'mistralai', 256000, NULL),
+    ('cohere/command-r-08-2024', 'Command R', 'open_router', 'cohere', 128000, 4096),
+    ('cohere/command-r-plus-08-2024', 'Command R Plus', 'open_router', 'cohere', 128000, 4096),
+    ('cohere/command-a-03-2025', 'Command A 03.2025', 'open_router', 'cohere', 256000, 8192),
+    ('cohere/command-a-reasoning-08-2025', 'Command A Reasoning 08.2025', 'open_router', 'cohere', 256000, 32768),
+    ('cohere/command-a-translate-08-2025', 'Command A Translate 08.2025', 'open_router', 'cohere', 128000, 8192),
+    ('cohere/command-a-vision-07-2025', 'Command A Vision 07.2025', 'open_router', 'cohere', 128000, 8192),
+    ('deepseek/deepseek-chat', 'DeepSeek Chat', 'open_router', 'deepseek', 64000, 2048),
+    ('deepseek/deepseek-r1-0528', 'DeepSeek R1', 'open_router', 'deepseek', 163840, 163840),
+    ('perplexity/sonar', 'Perplexity Sonar', 'open_router', 'perplexity', 127000, 8000),
+    ('perplexity/sonar-pro', 'Perplexity Sonar Pro', 'open_router', 'perplexity', 200000, 8000),
+    ('perplexity/sonar-deep-research', 'Perplexity Sonar Deep Research', 'open_router', 'perplexity', 128000, 16000),
+    ('perplexity/sonar-reasoning-pro', 'Sonar Reasoning Pro', 'open_router', 'perplexity', 128000, 8000),
+    ('nousresearch/hermes-3-llama-3.1-405b', 'Hermes 3 Llama 3.1 405B', 'open_router', 'nousresearch', 131000, 4096),
+    ('nousresearch/hermes-3-llama-3.1-70b', 'Hermes 3 Llama 3.1 70B', 'open_router', 'nousresearch', 12288, 12288),
+    ('openai/gpt-oss-120b', 'GPT OSS 120B', 'open_router', 'openai', 131072, 131072),
+    ('openai/gpt-oss-20b', 'GPT OSS 20B', 'open_router', 'openai', 131072, 32768),
+    ('amazon/nova-lite-v1', 'Amazon Nova Lite', 'open_router', 'amazon', 300000, 5120),
+    ('amazon/nova-micro-v1', 'Amazon Nova Micro', 'open_router', 'amazon', 128000, 5120),
+    ('amazon/nova-pro-v1', 'Amazon Nova Pro', 'open_router', 'amazon', 300000, 5120),
+    ('microsoft/wizardlm-2-8x22b', 'WizardLM 2 8x22B', 'open_router', 'microsoft', 65536, 4096),
+    ('microsoft/phi-4', 'Phi-4', 'open_router', 'microsoft', 16384, 16384),
+    ('gryphe/mythomax-l2-13b', 'MythoMax L2 13B', 'open_router', 'gryphe', 4096, 4096),
+    ('meta-llama/llama-4-scout', 'Llama 4 Scout', 'open_router', 'meta', 131072, 131072),
+    ('meta-llama/llama-4-maverick', 'Llama 4 Maverick', 'open_router', 'meta', 1048576, 1000000),
+    ('x-ai/grok-3', 'Grok 3', 'open_router', 'xai', 131072, 131072),
+    ('x-ai/grok-4', 'Grok 4', 'open_router', 'xai', 256000, 256000),
+    ('x-ai/grok-4-fast', 'Grok 4 Fast', 'open_router', 'xai', 2000000, 30000),
+    ('x-ai/grok-4.1-fast', 'Grok 4.1 Fast', 'open_router', 'xai', 2000000, 30000),
+    ('x-ai/grok-code-fast-1', 'Grok Code Fast 1', 'open_router', 'xai', 256000, 10000),
+    ('moonshotai/kimi-k2', 'Kimi K2', 'open_router', 'moonshotai', 131000, 131000),
+    ('qwen/qwen3-235b-a22b-thinking-2507', 'Qwen 3 235B Thinking', 'open_router', 'alibaba', 262144, 262144),
+    ('qwen/qwen3-coder', 'Qwen 3 Coder', 'open_router', 'alibaba', 262144, 262144),
+    -- Llama API models (creator: meta)
+    ('Llama-4-Scout-17B-16E-Instruct-FP8', 'Llama 4 Scout', 'llama_api', 'meta', 128000, 4028),
+    ('Llama-4-Maverick-17B-128E-Instruct-FP8', 'Llama 4 Maverick', 'llama_api', 'meta', 128000, 4028),
+    ('Llama-3.3-8B-Instruct', 'Llama 3.3 8B', 'llama_api', 'meta', 128000, 4028),
+    ('Llama-3.3-70B-Instruct', 'Llama 3.3 70B', 'llama_api', 'meta', 128000, 4028),
+    -- v0 models (creator: vercel)
+    ('v0-1.5-md', 'v0 1.5 MD', 'v0', 'vercel', 128000, 64000),
+    ('v0-1.5-lg', 'v0 1.5 LG', 'v0', 'vercel', 512000, 64000),
+    ('v0-1.0-md', 'v0 1.0 MD', 'v0', 'vercel', 128000, 64000)
+) AS models(model_slug, model_display_name, provider_name, creator_name, context_window, max_output_tokens)
+JOIN provider_ids p ON p."name" = models.provider_name
+JOIN creator_ids c ON c."name" = models.creator_name
+ON CONFLICT ("slug") DO NOTHING;
+
+-- Insert Costs (using CTEs to reference model IDs)
+WITH model_ids AS (
+    SELECT "id", "slug", "providerId" FROM "LlmModel"
+),
+provider_ids AS (
+    SELECT "id", "name" FROM "LlmProvider"
+)
+INSERT INTO "LlmModelCost" ("id", "createdAt", "updatedAt", "unit", "creditCost", "credentialProvider", "credentialId", "credentialType", "currency", "metadata", "llmModelId")
+SELECT
+    gen_random_uuid(),
+    CURRENT_TIMESTAMP,
+    CURRENT_TIMESTAMP,
+    'RUN'::"LlmCostUnit",
+    cost,
+    p."name",
+    NULL,
+    'api_key',
+    NULL,
+    '{}'::jsonb,
+    m."id"
+FROM (VALUES
+    -- OpenAI costs
+    ('o3-2025-04-16', 4),
+    ('o3-mini', 2),
+    ('o1', 16),
+    ('o1-mini', 4),
+    ('gpt-5.2-2025-12-11', 5),
+    ('gpt-5-2025-08-07', 2),
+    ('gpt-5.1-2025-11-13', 5),
+    ('gpt-5-mini-2025-08-07', 1),
+    ('gpt-5-nano-2025-08-07', 1),
+    ('gpt-5-chat-latest', 5),
+    ('gpt-4.1-2025-04-14', 2),
+    ('gpt-4.1-mini-2025-04-14', 1),
+    ('gpt-4o-mini', 1),
+    ('gpt-4o', 3),
+    ('gpt-4-turbo', 10),
+    -- Anthropic costs
+    ('claude-opus-4-6', 21),
+    ('claude-sonnet-4-6', 5),
+    ('claude-opus-4-1-20250805', 21),
+    ('claude-opus-4-20250514', 21),
+    ('claude-sonnet-4-20250514', 5),
+    ('claude-haiku-4-5-20251001', 4),
+    ('claude-opus-4-5-20251101', 14),
+    ('claude-sonnet-4-5-20250929', 9),
+    ('claude-3-haiku-20240307', 1),
+    -- AI/ML API costs
+    ('Qwen/Qwen2.5-72B-Instruct-Turbo', 1),
+    ('nvidia/llama-3.1-nemotron-70b-instruct', 1),
+    ('meta-llama/Llama-3.3-70B-Instruct-Turbo', 1),
+    ('meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo', 1),
+    ('meta-llama/Llama-3.2-3B-Instruct-Turbo', 1),
+    -- Groq costs
+    ('llama-3.3-70b-versatile', 1),
+    ('llama-3.1-8b-instant', 1),
+    -- Ollama costs
+    ('llama3.3', 1),
+    ('llama3.2', 1),
+    ('llama3', 1),
+    ('llama3.1:405b', 1),
+    ('dolphin-mistral:latest', 1),
+    -- OpenRouter costs
+    ('google/gemini-2.5-pro-preview-03-25', 4),
+    ('google/gemini-2.5-pro', 4),
+    ('google/gemini-3.1-pro-preview', 5),
+    ('google/gemini-3-flash-preview', 3),
+    ('google/gemini-3.1-flash-lite-preview', 1),
+    ('mistralai/mistral-nemo', 1),
+    ('mistralai/mistral-large-2512', 3),
+    ('mistralai/mistral-medium-3.1', 2),
+    ('mistralai/mistral-small-3.2-24b-instruct', 1),
+    ('mistralai/codestral-2508', 2),
+    ('cohere/command-r-08-2024', 1),
+    ('cohere/command-r-plus-08-2024', 3),
+    ('cohere/command-a-03-2025', 2),
+    ('cohere/command-a-reasoning-08-2025', 3),
+    ('cohere/command-a-translate-08-2025', 1),
+    ('cohere/command-a-vision-07-2025', 2),
+    ('deepseek/deepseek-chat', 2),
+    ('perplexity/sonar', 1),
+    ('perplexity/sonar-pro', 5),
+    ('perplexity/sonar-deep-research', 10),
+    ('perplexity/sonar-reasoning-pro', 5),
+    ('nousresearch/hermes-3-llama-3.1-405b', 1),
+    ('nousresearch/hermes-3-llama-3.1-70b', 1),
+    ('amazon/nova-lite-v1', 1),
+    ('amazon/nova-micro-v1', 1),
+    ('amazon/nova-pro-v1', 1),
+    ('microsoft/wizardlm-2-8x22b', 1),
+    ('microsoft/phi-4', 1),
+    ('gryphe/mythomax-l2-13b', 1),
+    ('meta-llama/llama-4-scout', 1),
+    ('meta-llama/llama-4-maverick', 1),
+    ('x-ai/grok-3', 5),
+    ('x-ai/grok-4', 9),
+    ('x-ai/grok-4-fast', 1),
+    ('x-ai/grok-4.1-fast', 1),
+    ('x-ai/grok-code-fast-1', 1),
+    ('moonshotai/kimi-k2', 1),
+    ('qwen/qwen3-235b-a22b-thinking-2507', 1),
+    ('qwen/qwen3-coder', 9),
+    ('google/gemini-2.5-flash', 1),
+    ('google/gemini-2.0-flash-001', 1),
+    ('google/gemini-2.5-flash-lite-preview-06-17', 1),
+    ('google/gemini-2.0-flash-lite-001', 1),
+    ('deepseek/deepseek-r1-0528', 1),
+    ('openai/gpt-oss-120b', 1),
+    ('openai/gpt-oss-20b', 1),
+    -- Llama API costs
+    ('Llama-4-Scout-17B-16E-Instruct-FP8', 1),
+    ('Llama-4-Maverick-17B-128E-Instruct-FP8', 1),
+    ('Llama-3.3-8B-Instruct', 1),
+    ('Llama-3.3-70B-Instruct', 1),
+    -- v0 costs
+    ('v0-1.5-md', 1),
+    ('v0-1.5-lg', 2),
+    ('v0-1.0-md', 1)
+) AS costs(model_slug, cost)
+JOIN model_ids m ON m."slug" = costs.model_slug
+JOIN provider_ids p ON p."id" = m."providerId"
+ON CONFLICT ("llmModelId", "credentialProvider", "unit") WHERE "credentialId" IS NULL DO NOTHING;
+

autogpt_platform/backend/schema.prisma

@@ -1301,3 +1301,164 @@ model OAuthRefreshToken {
   @@index([userId, applicationId])
   @@index([expiresAt]) // For cleanup
 }
+
+// ============================================================================
+// LLM Registry Models
+// ============================================================================
+
+enum LlmCostUnit {
+  RUN
+  TOKENS
+}
+
+model LlmProvider {
+  id        String   @id @default(uuid())
+  createdAt DateTime @default(now())
+  updatedAt DateTime @updatedAt
+
+  name        String @unique
+  displayName String
+  description String?
+
+  defaultCredentialProvider String?
+  defaultCredentialId       String?
+  defaultCredentialType     String?
+
+  metadata Json @default("{}")
+
+  Models LlmModel[]
+
+}
+
+model LlmModel {
+  id        String   @id @default(uuid())
+  createdAt DateTime @default(now())
+  updatedAt DateTime @updatedAt
+
+  slug        String @unique
+  displayName String
+  description String?
+
+  providerId String
+  Provider   LlmProvider @relation(fields: [providerId], references: [id], onDelete: Restrict)
+
+  // Creator is the organization that created/trained the model (e.g., OpenAI, Meta)
+  // This is distinct from the provider who hosts/serves the model (e.g., OpenRouter)
+  creatorId String?
+  Creator   LlmModelCreator? @relation(fields: [creatorId], references: [id], onDelete: SetNull)
+
+  contextWindow   Int
+  maxOutputTokens Int?
+  priceTier       Int     @default(1) // 1=cheapest, 2=medium, 3=expensive (DB constraint: 1-3)
+  isEnabled       Boolean @default(true)
+  isRecommended   Boolean @default(false)
+
+  // Model-specific capabilities
+  // These vary per model even within the same provider (e.g., Hugging Face)
+  // Default to false for safety - partially-seeded rows should not be assumed capable
+  supportsTools            Boolean @default(false)
+  supportsJsonOutput       Boolean @default(false)
+  supportsReasoning        Boolean @default(false)
+  supportsParallelToolCalls Boolean @default(false)
+
+  capabilities Json @default("{}")
+  metadata     Json @default("{}")
+
+  Costs             LlmModelCost[]
+  SourceMigrations  LlmModelMigration[] @relation("SourceMigrations")
+  TargetMigrations  LlmModelMigration[] @relation("TargetMigrations")
+
+  @@index([providerId, isEnabled])
+  @@index([creatorId])
+  // Note: slug already has @unique which creates an implicit index
+}
+
+model LlmModelCost {
+  id        String      @id @default(uuid())
+  createdAt DateTime    @default(now())
+  updatedAt DateTime    @updatedAt
+  unit      LlmCostUnit @default(RUN)
+
+  creditCost Int // DB constraint: >= 0
+
+  // Provider identifier (e.g., "openai", "anthropic", "openrouter")
+  // Used to determine which credential system provides the API key.
+  // Allows different pricing for:
+  // - Default provider costs (WHERE credentialId IS NULL)
+  // - User's own API key costs (WHERE credentialId IS NOT NULL)
+  credentialProvider String
+  credentialId       String?
+  credentialType     String?
+  currency           String?
+
+  metadata Json @default("{}")
+
+  llmModelId String
+  Model      LlmModel @relation(fields: [llmModelId], references: [id], onDelete: Cascade)
+
+  // Note: Unique constraints are implemented as partial indexes in migration SQL:
+  // - One for default costs (WHERE credentialId IS NULL)
+  // - One for credential-specific costs (WHERE credentialId IS NOT NULL)
+  // This allows both provider-level defaults and credential-specific overrides
+}
+
+model LlmModelCreator {
+  id        String   @id @default(uuid())
+  createdAt DateTime @default(now())
+  updatedAt DateTime @updatedAt
+
+  name        String  @unique // e.g., "openai", "anthropic", "meta"
+  displayName String  // e.g., "OpenAI", "Anthropic", "Meta"
+  description String?
+  websiteUrl  String? // Link to creator's website
+  logoUrl     String? // URL to creator's logo
+
+  metadata Json @default("{}")
+
+  Models LlmModel[]
+
+}
+
+model LlmModelMigration {
+  id        String   @id @default(uuid())
+  createdAt DateTime @default(now())
+  updatedAt DateTime @updatedAt
+
+  sourceModelSlug String // The original model that was disabled
+  targetModelSlug String // The model workflows were migrated to
+  reason          String? // Why the migration happened (e.g., "Provider outage")
+
+  // FK constraints ensure slugs reference valid models
+  SourceModel LlmModel @relation("SourceMigrations", fields: [sourceModelSlug], references: [slug], onDelete: Restrict)
+  TargetModel LlmModel @relation("TargetMigrations", fields: [targetModelSlug], references: [slug], onDelete: Restrict)
+
+  // Track affected nodes as JSON array of node IDs
+  // Format: ["node-uuid-1", "node-uuid-2", ...]
+  migratedNodeIds Json @default("[]")
+  nodeCount       Int  // Number of nodes migrated (DB constraint: >= 0)
+
+  // Custom pricing override for migrated workflows during the migration period.
+  // Use case: When migrating users from an expensive model (e.g., GPT-4) to a cheaper
+  // one (e.g., GPT-3.5), you may want to temporarily maintain the original pricing
+  // to avoid billing surprises, or offer a discount during the transition.
+  //
+  // IMPORTANT: This field is intended for integration with the billing system.
+  // When billing calculates costs for nodes affected by this migration, it should
+  // check if customCreditCost is set and use it instead of the target model's cost.
+  // If null, the target model's normal cost applies.
+  //
+  // TODO: Integrate with billing system to apply this override during cost calculation.
+  // LIMITATION: This is a simple Int and doesn't distinguish RUN vs TOKENS pricing.
+  // For token-priced models, this may be ambiguous. Consider migrating to a relation
+  // with LlmModelCost or a dedicated override model in a follow-up PR.
+  customCreditCost Int? // DB constraint: >= 0 when not null
+
+  // Revert tracking
+  isReverted Boolean   @default(false)
+  revertedAt DateTime?
+
+  // Note: Partial unique index in migration SQL prevents multiple active migrations per source:
+  // UNIQUE (sourceModelSlug) WHERE isReverted = false
+  @@index([targetModelSlug])
+  @@index([sourceModelSlug, isReverted]) // Composite index for active migration queries
+}

autogpt_platform/frontend/src/app/api/openapi.json

@@ -4603,6 +4603,74 @@
         }
       }
     },
+    "/api/llm/models": {
+      "get": {
+        "tags": ["v2", "llm", "llm"],
+        "summary": "List Models",
+        "description": "List all LLM models available to users.\n\nReturns models from the in-memory registry cache.\nUse enabled_only=true to filter to only enabled models (default).",
+        "operationId": "getV2ListModels",
+        "security": [{ "HTTPBearerJWT": [] }],
+        "parameters": [
+          {
+            "name": "enabled_only",
+            "in": "query",
+            "required": false,
+            "schema": {
+              "type": "boolean",
+              "description": "Only return enabled models",
+              "default": true,
+              "title": "Enabled Only"
+            },
+            "description": "Only return enabled models"
+          }
+        ],
+        "responses": {
+          "200": {
+            "description": "Successful Response",
+            "content": {
+              "application/json": {
+                "schema": { "$ref": "#/components/schemas/LlmModelsResponse" }
+              }
+            }
+          },
+          "401": {
+            "$ref": "#/components/responses/HTTP401NotAuthenticatedError"
+          },
+          "422": {
+            "description": "Validation Error",
+            "content": {
+              "application/json": {
+                "schema": { "$ref": "#/components/schemas/HTTPValidationError" }
+              }
+            }
+          }
+        }
+      }
+    },
+    "/api/llm/providers": {
+      "get": {
+        "tags": ["v2", "llm", "llm"],
+        "summary": "List Providers",
+        "description": "List all LLM providers with their enabled models.\n\nGroups enabled models by provider from the in-memory registry.",
+        "operationId": "getV2ListProviders",
+        "responses": {
+          "200": {
+            "description": "Successful Response",
+            "content": {
+              "application/json": {
+                "schema": {
+                  "$ref": "#/components/schemas/LlmProvidersResponse"
+                }
+              }
+            }
+          },
+          "401": {
+            "$ref": "#/components/responses/HTTP401NotAuthenticatedError"
+          }
+        },
+        "security": [{ "HTTPBearerJWT": [] }]
+      }
+    },
     "/api/mcp/discover-tools": {
       "post": {
         "tags": ["v2", "mcp", "mcp"],
@@ -10257,6 +10325,159 @@
         "title": "ListSessionsResponse",
         "description": "Response model for listing chat sessions."
       },
+      "LlmModel": {
+        "properties": {
+          "slug": { "type": "string", "title": "Slug" },
+          "display_name": { "type": "string", "title": "Display Name" },
+          "description": {
+            "anyOf": [{ "type": "string" }, { "type": "null" }],
+            "title": "Description"
+          },
+          "provider_name": { "type": "string", "title": "Provider Name" },
+          "creator": {
+            "anyOf": [
+              { "$ref": "#/components/schemas/LlmModelCreator" },
+              { "type": "null" }
+            ]
+          },
+          "context_window": { "type": "integer", "title": "Context Window" },
+          "max_output_tokens": {
+            "anyOf": [{ "type": "integer" }, { "type": "null" }],
+            "title": "Max Output Tokens"
+          },
+          "price_tier": { "type": "integer", "title": "Price Tier" },
+          "is_enabled": {
+            "type": "boolean",
+            "title": "Is Enabled",
+            "default": true
+          },
+          "is_recommended": {
+            "type": "boolean",
+            "title": "Is Recommended",
+            "default": false
+          },
+          "capabilities": {
+            "additionalProperties": true,
+            "type": "object",
+            "title": "Capabilities"
+          },
+          "costs": {
+            "items": { "$ref": "#/components/schemas/LlmModelCost" },
+            "type": "array",
+            "title": "Costs"
+          }
+        },
+        "type": "object",
+        "required": [
+          "slug",
+          "display_name",
+          "provider_name",
+          "context_window",
+          "price_tier"
+        ],
+        "title": "LlmModel",
+        "description": "Public-facing LLM model information."
+      },
+      "LlmModelCost": {
+        "properties": {
+          "unit": { "type": "string", "title": "Unit" },
+          "credit_cost": {
+            "type": "integer",
+            "minimum": 0.0,
+            "title": "Credit Cost"
+          },
+          "credential_provider": {
+            "type": "string",
+            "title": "Credential Provider"
+          },
+          "credential_id": {
+            "anyOf": [{ "type": "string" }, { "type": "null" }],
+            "title": "Credential Id"
+          },
+          "credential_type": {
+            "anyOf": [{ "type": "string" }, { "type": "null" }],
+            "title": "Credential Type"
+          },
+          "currency": {
+            "anyOf": [{ "type": "string" }, { "type": "null" }],
+            "title": "Currency"
+          },
+          "metadata": {
+            "additionalProperties": true,
+            "type": "object",
+            "title": "Metadata"
+          }
+        },
+        "type": "object",
+        "required": ["unit", "credit_cost", "credential_provider"],
+        "title": "LlmModelCost",
+        "description": "Cost configuration for an LLM model."
+      },
+      "LlmModelCreator": {
+        "properties": {
+          "id": { "type": "string", "title": "Id" },
+          "name": { "type": "string", "title": "Name" },
+          "display_name": { "type": "string", "title": "Display Name" },
+          "description": {
+            "anyOf": [{ "type": "string" }, { "type": "null" }],
+            "title": "Description"
+          },
+          "website_url": {
+            "anyOf": [{ "type": "string" }, { "type": "null" }],
+            "title": "Website Url"
+          },
+          "logo_url": {
+            "anyOf": [{ "type": "string" }, { "type": "null" }],
+            "title": "Logo Url"
+          }
+        },
+        "type": "object",
+        "required": ["id", "name", "display_name"],
+        "title": "LlmModelCreator",
+        "description": "Represents the organization that created/trained the model."
+      },
+      "LlmModelsResponse": {
+        "properties": {
+          "models": {
+            "items": { "$ref": "#/components/schemas/LlmModel" },
+            "type": "array",
+            "title": "Models"
+          },
+          "total": { "type": "integer", "title": "Total" }
+        },
+        "type": "object",
+        "required": ["models", "total"],
+        "title": "LlmModelsResponse",
+        "description": "Response for GET /llm/models."
+      },
+      "LlmProvider": {
+        "properties": {
+          "name": { "type": "string", "title": "Name" },
+          "display_name": { "type": "string", "title": "Display Name" },
+          "models": {
+            "items": { "$ref": "#/components/schemas/LlmModel" },
+            "type": "array",
+            "title": "Models"
+          }
+        },
+        "type": "object",
+        "required": ["name", "display_name"],
+        "title": "LlmProvider",
+        "description": "Provider with its enabled models."
+      },
+      "LlmProvidersResponse": {
+        "properties": {
+          "providers": {
+            "items": { "$ref": "#/components/schemas/LlmProvider" },
+            "type": "array",
+            "title": "Providers"
+          }
+        },
+        "type": "object",
+        "required": ["providers"],
+        "title": "LlmProvidersResponse",
+        "description": "Response for GET /llm/providers."
+      },
       "LogRawMetricRequest": {
         "properties": {
           "metric_name": {