- Follow-up fix to #10167
- Resolves#10228
### Changes 🏗️
- Don't assume `block.input_schema.jsonschema()["required"]` exists
- Unbreak handling of `webhook_type` in
`BaseWebhooksManager.get_manual_webhook(..)`
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- Create an agent with a Generic Webhook Trigger block; go to it in the
Library
- [x] -> `/library/agents/[id]` loads normally
AIImageEditorBlock was not able to accept an image from AgentFileInput
or FileStore block.
### Changes 🏗️
* Add support for image loading for the image editor block:
<img width="1081" alt="Screenshot 2025-06-23 at 10 28 45 AM"
src="https://github.com/user-attachments/assets/ac3fea91-9503-4894-bbe3-2dc3c5649a39"
/>
* Avoid rendering a relative path image file.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Run AiImageEditor block using AgentFileInput or FileStore block.
This pull request adds support for setting up (webhook-)triggered agents
in the Library. It contains changes throughout the entire stack to make
everything work in the various phases of a triggered agent's lifecycle:
setup, execution, updates, deletion.
Setting up agents with webhook triggers was previously only possible in
the Builder, limiting their use to the agent's creator only. To make it
work in the Library, this change uses the previously introduced
`AgentPreset` to store information on, instead of on the graph's nodes
to which only a graph's creator has access.
- Initial ticket: #10111
- Builds on #9786
### Changes 🏗️
Frontend:
- Amend the Library's `AgentRunDraftView` to handle creating and editing
Presets
- Add `hideIfSingleCredentialAvailable` parameter to `CredentialsInput`
- Add multi-select support to `TypeBasedInput`
- Add Presets section to `AgentRunsSelectorList`
- Amend `AgentRunSummaryCard` for use for Presets
- Add `AgentStatusChip` to display general agent status (for now: Active
/ Inactive / Error)
- Add Preset loading logic and create/update/delete handlers logic to
`AgentRunsPage`
- Rename `IconClose` to `IconCross`
API:
- Add `LibraryAgent` properties `has_external_trigger`,
`trigger_setup_info`, `credentials_input_schema`
- Add `POST /library/agents/{library_agent_id}/setup_trigger` endpoint
- Remove redundant parameters from `POST
/library/presets/{preset_id}/execute` endpoint
Backend:
- Add `POST /library/agents/{library_agent_id}/setup_trigger` endpoint
- Extract non-node-related logic from `on_node_activate` into
`setup_webhook_for_block`
- Add webhook-related logic to `update_preset` and `delete_preset`
endpoints
- Amend webhook infrastructure to work with AgentPresets
- Add preset trigger support to webhook ingress endpoint
- Amend executor stack to work with passed-in node input
(`nodes_input_masks`, generalized from `node_credentials_input_map`)
- Amend graph validation to work with passed-in node input
- Add `AgentPreset`->`IntegrationWebhook` relation
- Add `WebhookWithRelations` model
- Change behavior of `BaseWebhooksManager.get_manual_webhook(..)` to
avoid unnecessary changes of the webhook URL: ignore `events` to find
matching webhook, and update `events` if necessary.
- Fix & improve `AgentPreset` API, models, and DB logic
- Add `isDeleted` filter to get/list queries
- Add `user_id` attribute to `LibraryAgentPreset` model
- Add separate `credentials` property to `LibraryAgentPreset` model
- Fix `library_db.update_preset(..)` replacement of existing
`InputPresets`
- Make `library_db.update_preset(..)` more usage-friendly with separate
parameters for updateable properties
- Add `user_id` checks to various DB functions
- Fix error handling in various endpoints
- Fix cache race condition on `load_webhook_managers()`
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- Test existing functionality
- [x] Auto-setup and -teardown of webhooks on save in the builder still
works
- [x] Running an agent normally from the Library still works
- Test new functionality
- [x] Setting up a trigger in the Library
- [x] Updating a trigger in the Library
- [x] Disabling and re-enabling a trigger in the Library
- [x] Deleting a trigger in the Library
- [x] Triggers set up in the Library result in a new run when the
webhook receives a payload
This pull request sets up and configures Orval for API client
generation. It automates the process of creating TypeScript clients from
the backend's OpenAPI specification, improving development efficiency
and reducing manual code maintenance.
### Changes 🏗️
- Configures Orval with a new configuration file (`orval.config.ts`).
- Adds scripts to `package.json` for fetching the OpenAPI spec and
generating the API client.
- Implements a custom mutator for handling authentication.
- Adds API client generation as a step in the CI workflow.
- Adds `.gitignore` entry for generated API client files.
- Adds a security middleware to prevent caching of sensitive data.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Verified that the API client is generated correctly.
- [x] Confirmed that the custom mutator is functioning as expected for
authentication.
- [x] Ensured that the new CI workflow step for API client generation is
successful.
- [x] Tested generated API calls
#### For configuration changes:
- [x] `.env.example` is updated or already compatible with my changes
- [ ] `docker-compose.yml` is updated or already compatible with my
changes
- [x] I have included a list of my configuration changes in the PR
description (under **Changes**)
Since auto conversion is applied before merging nested input in the
block, it breaks the auto conversion break.
### Changes 🏗️
* Enabling auto-type conversion on block input schema mismatch for
nested input
* Add batching feature for `CreateListBlock`
* Increase default max_token size for LLM call
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Run `AIStructuredResponseGeneratorBlock` with non-string prompt
value (should be auto-converted).
### Changes 🏗️
Add cost calculation for Apollo integration.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Run Apollo block Search People & Organizations Block.
### Why? 🤔
<!-- Clearly explain the need for these changes: -->
We need to prevent sensitive data (authentication tokens, API
keys, user credentials, personal information) from being cached by
browsers and proxies. Following the principle of "secure by
default", we're switching from a deny list to an allow list
approach for cache control.
### Changes 🛠️
<!-- Concisely describe all of the changes made in this pull
request: -->
- **Refactored cache control middleware from deny list to allow
list approach**
- By default, ALL endpoints now have `Cache-Control: no-store,
no-cache, must-revalidate, private` headers
- Only explicitly allowed paths (static assets, health checks,
public store pages) can be cached
- This ensures new endpoints are automatically protected without
developers having to remember to add them to a list
- **Updated `SecurityHeadersMiddleware` in
`/backend/backend/server/middleware/security.py`**
- Renamed `SENSITIVE_PATHS` to `CACHEABLE_PATHS`
- Inverted the logic in `is_cacheable_path()` method
- Cache control headers are now applied to all paths NOT in the
allow list
- **Updated test suite to match new behavior**
- Tests now verify that most endpoints have cache control
headers by default
- Tests verify that only allowed paths (static assets, health
endpoints, etc.) can be cached
- **Updated documentation in `CLAUDE.md`**
- Documented the new allow list approach
- Added instructions for developers on how to allow caching for
new endpoints
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Test modified endpoints work still
- [x] Test modified endpoints correctly have no cache rules
---------
Co-authored-by: Swifty <craigswift13@gmail.com>
Main issues:
* `AIStructuredResponseGeneratorBlock` is not able to produce a list of
objects.
* `SmartDecisionBlock` is not able to call tools with some optional
inputs.
### Changes 🏗️
* Allow persisting `null` / `None` value as execution output.
* Provide `multiple_tool_calls` option for `SmartDecisionBlock`.
* Provide `list_result` option for `AIStructuredResponseGeneratorBlock`
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Run `SmartDecisionBlock` & `AIStructuredResponseGeneratorBlock`
This PR introduces a custom function for generating unique operation IDs
for OpenAPI specifications to improve auto-generated client code
quality.
## Why This Change?
**Better Auto-Generated Clients**: Default FastAPI operation IDs create
unclear method names in generated clients. Our custom generator produces
clean, readable operation IDs that translate to intuitive method names.
- **Before**: `get_items_api_v1_items_get` → unclear generated methods
- **After**: `get_users_list` → clean, descriptive method names
## Changes
- ✨ **Added**: `custom_generate_unique_id` utility function
- Generates IDs using pattern: `{method}_{tag}_{summary}`
- Ensures uniqueness and readability
- 🔧 **Updated**: FastAPI app configuration to use custom generator
## Checklist
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] OpenAPI docs reflect new operation ID format
- [x] Tested various HTTP methods, tags, and summaries
- [x] Verified app startup functionality
- [x] Validated improved client generation output
Current Apollo blocks only work with keywords; the huge number of list
filter fields doesn't work because it's passing the wrong GET parameter
(missing `[]`).
### Changes 🏗️
Change the GET request to a POST request for Apollo.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Run SearchPeopleBlock with title filter
## Description
Added the `graph_id` parameter to the stop execution endpoint path
(`/graphs/{graph_id}/executions/{graph_exec_id}/stop`) to fix client
generation from Openapi spec error.
## Problem
The client generation was failing due to missing path parameter
definition for `graph_id` in the stop execution endpoint.
<img width="1412" alt="Screenshot 2025-06-19 at 9 20 17 AM"
src="https://github.com/user-attachments/assets/aa1667d3-05be-48c6-975b-84473830ac03"
/>
## Solution
Added `graph_id` as a path parameter while maintaining the existing
functionality.
## Testing
- [x] Verified OpenAPI client generation works without errors
- [x] Confirmed endpoint functionality remains unchanged
- [x] Tested API calls maintain backward compatibility
Request on block execution can be throttled, and requests between
services can sometimes break. The scope of this PR is to add an
appropriate retry on those.
### Changes 🏗️
* Block request retry: Retry on throttled status code only (504, 429,
etc).
* RPC request retry: Retry connection issues (ConnectError, Timeout,
etc).
* Truncate logging on executor/utils.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Manual graph execution
- Follow-up fix for #9786
A change to a DB statement introduced in #9786 turns out to be breaking.
Apparently `connect` can't just be used for *some* relations: if it is
used, it must be used for *all* relations created by the statement.
### Changes 🏗️
- Fix broken DB statement in `add_store_agent_to_library(..)`
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Add store agent to library
Co-authored-by: Swifty <craigswift13@gmail.com>
This change introduced async execution for blocks and the execution
engine. Paralellism will be achieved through a single process
asynchronous execution instead of process concurrency.
### Changes 🏗️
* Support async execution for the graph executor
* Removed process creation for node execution
* Update all blocks to support async executions
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Manual graph executions, tested many of the impacted blocks.
<!-- Clearly explain the need for these changes: -->
Doing the CASA Audit and this is something to check
### Changes 🏗️
- limits APIs to use their specific endpoints
- use expected trusted sources for each block and requests call
- Use cryptographically valid string comparisons
- Don't log secrets
<!-- Concisely describe all of the changes made in this pull request:
-->
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Testing in dev branch once merged
---------
Co-authored-by: Swifty <craigswift13@gmail.com>
<!-- Clearly explain the need for these changes: -->
## Background & Summary of Changes
If a user has a single invalid Agent in their Library (i.e one with a
Block which doesn't exist) currently the Blocks menu does not return any
Agent results.
Valid agents should still load even when some stored graphs are
malformed.
Graphs which are malformed should just be skipped rather than breaking
the entire process, this PR implements that fix, unblocking users with a
malformed Agent in their Library (me!).
## Testing
I have tested this PR in the dev deployment (where I have this issue on
my account) and have confirmed that Agents now show up in the list:
| Before this Change | After this Change |
| ------------------ | ----------------- |
| 
| 
|
## Changes 🏗️
- Validate each graph’s serialization in get_graphs and skip any that
raise an exception
- Added error logging for invalid graphs
## Checklist 📋
For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] poetry run format
- [ ] poetry run test
For configuration changes:
- [x] .env.example is updated or already compatible with my changes
- [x] docker-compose.yml is updated or already compatible with my
changes
- [x] I have included a list of my configuration changes in the PR
description (under Changes)
Fixes [OPEN-2461: Loading a Library Agent with an invalid block causes
all Library Agent Loading to fail in Builder Blocks
Menu](https://linear.app/autogpt/issue/OPEN-2461/loading-a-library-agent-with-an-invalid-block-causes-all-library-agent)
<img width="1410" alt="image"
src="https://github.com/user-attachments/assets/bce407a2-96a1-42e9-9772-b49b8f20886c"
/>
### Changes 🏗️
Add the missing `send execution update` command on completed/update
status change for the node execution.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Screenshot attached
### Changes 🏗️
This simply adds "fal-ai/veo3" to the ``FalModel`` in the
``ai_video_generator.py`` file
Oh i also set it so veo3 also always generates videos with audio so
``generate_audio=True`` is set to true if veo3 is selected
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Test the veo3 model via Fal.ai and it should work.
### Changes 🏗️
Today openAI dropped the prices of the o3 model so this simply drops the
price from 7 to 4
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Run the platform with the new price, check the O3 model in the ai
text generator block and see its cheaper to use
Allowing depth-first execution will unlock faster processing latency and
a better sense of progress.
<img width="950" alt="image"
src="https://github.com/user-attachments/assets/e2a0e11a-8bc5-4a65-a10d-b5b6c6383354"
/>
### Changes 🏗️
* Prioritize adding a new execution over processing execution output
* Make sure to enqueue each node once when processing output instead of
draining a single node and move on.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Run company follower count finder agent.
---------
Co-authored-by: Swifty <craigswift13@gmail.com>
## Changes
- log helpful hints when metrics fail to record
- clarify API key errors in v1 router
- improve Postmark unsubscribe and webhook logs
- surface actionable feedback across integrations and store APIs
- handle Otto proxy failures with guidance
## Checklist
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan
There are a few UI bugs on the builder that this PR addresses.
<img width="554" alt="image"
src="https://github.com/user-attachments/assets/1be70197-de7e-40fe-ab11-405c145e763d"
/>
### Changes 🏗️
Fix these UI issues:
* (screenshot attached above) Key-value input width was unintentionally
maxed out due to a stale CSS rule.
* When multiple executions within the same node are running, we pick the
latest status, making one running and one completed execution displayed
as completed.
* No balance errors were executed, only displayed while at least one
node execution was triggered, while this can be done directly when the
execution request is triggered.
* Run & Stop button glitch: it's still showing as stopped when the graph
is still running, this is due to way the UI code tracks execution in the
node-level, instead of graph level.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Manual tests on the described behaviours.
This pull request introduces a comprehensive backend testing guide and
adds new tests for analytics logging and various API endpoints, focusing
on snapshot testing. It also includes corresponding snapshot files for
these tests. Below are the most significant changes:
### Documentation Updates:
* Added a detailed `TESTING.md` file to the backend, providing a guide
for running tests, snapshot testing, writing API route tests, and best
practices. It includes examples for mocking, fixtures, and CI/CD
integration.
### Analytics Logging Tests:
* Implemented tests for logging raw metrics and analytics in
`analytics_test.py`, covering success scenarios, various input values,
invalid requests, and complex nested data. These tests utilize snapshot
testing for response validation.
* Added snapshot files for analytics logging tests, including responses
for success cases, various metric values, and complex analytics data.
[[1]](diffhunk://#diff-654bc5aa1951008ec5c110a702279ef58709ee455ba049b9fa825fa60f7e3869R1-R3)
[[2]](diffhunk://#diff-e0a434b107abc71aeffb7d7989dbfd8f466b5e53f8dea25a87937ec1b885b122R1-R3)
[[3]](diffhunk://#diff-dd0bc0b72264de1a0c0d3bd0c54ad656061317f425e4de461018ca51a19171a0R1-R3)
[[4]](diffhunk://#diff-63af007073db553d04988544af46930458a768544cabd08412265e0818320d11R1-R30)
### Snapshot Files for API Endpoints:
* Added snapshot files for various API endpoint tests, such as:
- Graph-related operations (`graphs_get_single_response`,
`graphs_get_all_response`, `blocks_get_all_response`).
[[1]](diffhunk://#diff-b25dba271606530cfa428c00073d7e016184a7bb22166148ab1726b3e113dda8R1-R29)
[[2]](diffhunk://#diff-1054e58ec3094715660f55bfba1676d65b6833a81a91a08e90ad57922444d056R1-R31)
[[3]](diffhunk://#diff-cfd403ab6f3efc89188acaf993d85e6f792108d1740c7e7149eb05efb73d918dR1-R14)
- User-related operations (`auth_get_or_create_user_response`,
`auth_update_email_response`).
[[1]](diffhunk://#diff-49e65ab1eb6af4d0163a6c54ed10be621ce7336b2ab5d47d47679bfaefdb7059R1-R5)
[[2]](diffhunk://#diff-ac1216f96878bd4356454c317473654d5d5c7c180125663b80b0b45aa5ab52cbR1-R3)
- Credit-related operations (`credits_get_balance_response`,
`credits_get_auto_top_up_response`, `credits_top_up_request_response`).
[[1]](diffhunk://#diff-189488f8da5be74d80ac3fb7f84f1039a408573184293e9ba2e321d535c57cddR1-R3)
[[2]](diffhunk://#diff-ba3c4a6853793cbed24030cdccedf966d71913451ef8eb4b2c4f426ef18ed87aR1-R4)
[[3]](diffhunk://#diff-43d7daa0c82070a9b6aee88a774add8e87533e630bbccbac5a838b7a7ae56a75R1-R3)
- Graph execution and deletion (`blocks_execute_response`,
`graphs_delete_response`).
[[1]](diffhunk://#diff-a2ade7d646ad85a2801e7ff39799a925a612548a1cdd0ed99b44dd870d1465b5R1-R12)
[[2]](diffhunk://#diff-c0d1cd0a8499ee175ce3007c3a87ba5f3235ce02d38ce837560b36a44fdc4a22R1-R3)##
Summary
- add pytest-snapshot to backend dev requirements
- snapshot server route response JSONs
- mention how to update stored snapshots
## Testing
- `poetry run format`
- `poetry run test`
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] run poetry run test
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Nicholas Tindle <nicholas.tindle@agpt.co>
- Part of #9307
- ❗ Blocks #9541
### Changes 🏗️
Backend:
- Fix+improve `GET /library/presets` (`list_presets`) endpoint
- Fix pagination
- Add `graph_id` filter parameter
- Allow partial preset updates: `PUT /presets/{preset_id}` -> `PATCH
/presets/{preset_id}`
- Allow creating preset from graph execution through `POST /presets`
- Clean up models & DB functions
- Split `upsert_preset` into `create_preset` + `update_preset`
- Add `LibraryAgentPresetUpdatable`
- Replace `CreateLibraryAgentPresetRequest` with
`LibraryAgentPresetCreatable`
- Use `LibraryAgentPresetCreatable` as base class for
`LibraryAgentPreset`
- Remove redundant `set_is_deleted_for_library_agent(..)`
- Improve log statements
- Improve DB statements (e.g. by using unique keys where possible)
Frontend:
- Add timestamp parsing logic to library agent preset endpoints
- Brand `LibraryAgentPreset.id` + references
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] CI green
- Since these changes don't affect existing front-end functionality, no
additional testing is needed.
**Goal**: Allow parallel runs within a single node. Currently, we
prevent this to avoid unexpected ordering of the execution.
### Changes 🏗️
#### Executor changes
We decoupled the node execution output processing, which is responsible
for deciding the next executions from the node executor code.
Currently, `execute_node` does two big things:
* Runs the block’s execute(...) (which yields outputs).
* immediately enqueues the next nodes based on those outputs.
This PR makes:
* execute_node(node_exec) -> stream of (output_name, data). That purely
runs the block and yields each output as soon as it’s available.
* Move _enqueue_next_nodes into the graph executor. So the next
execution is handled serially by the graph executor to avoid concurrency
issues.
#### UI changes
The change on the executor also fixes the behavior of the execution
update to the UI We will report the execution output to the UI as soon
as it is available, not when the node execution is fully completed.
This, however, broke the bread calculation logic that assumes each
execution update will never overlap. So the change in this PR makes the
bead calculation take the overlap / duplicated execution update into
account, and simplify the overall calculation logic.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Execute this agent and observe its concurrency ordering
<img width="1424" alt="image"
src="https://github.com/user-attachments/assets/0fe8259f-9091-4ecc-b824-ce8e8819c2d2"
/>
Currently, the get_graph function, with no graph version specifier will
try to fetch the latest version, and when the graph is not owned and the
latest version is not available for listing, it will return `None`
instead of picking the latest graph version available on the store.
### Changes 🏗️
Instead of using the latest graph.version to fetch the store listing,
don't provide any version filter at all and pick up whatever available
version in the store.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] CI, existing tests
Suppose we have pint with list[list[int]] type, and we want directly
insert the a new value inside the first index of the first list e.g:
list[0][0] = X through a dynamic pin, this will be translated into
list_$_0_$_0, and the system does not currently support this.
### Changes 🏗️
Add support for nested dynamic pins for list, object, and dict.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] lots of unit tests
- [x] Tried inserting the value directly on the `value` nested field on
Google Sheets Write block.
<img width="371" alt="image"
src="https://github.com/user-attachments/assets/0a5e7213-b0e0-4fce-9e89-b39f7a583582"
/>
Now, SendWebRequestBlock can upload files. To make this work, we also
need to improve the UI rendering on the key-value pair input so that it
can also render media/file upload.
### Changes 🏗️
* Add file multipart upload support for SendWebRequestBlock
* Improve key-value input UI rendering to allow rendering any types as a
normal input block (it was only string & number).
<img width="381" alt="image"
src="https://github.com/user-attachments/assets/b41d778d-8f9d-4aec-95b6-0b32bef50e89"
/>
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Test running http request block, othe key-value pair input block
<!-- Clearly explain the need for these changes: -->
This PR adds a new internal block, **AI Image Editor**, which enables
**text-based image editing** via BlackForest Labs’ Flux Kontext models
on Replicate. This block allows users to input a prompt and optionally a
reference image, and returns a transformed image URL. It supports two
model variants (Pro and Max), with different cost tiers. This
functionality will enhance multimedia capabilities across internal agent
workflows and support richer AI-powered image manipulation.
---
### Changes 🏗️
* Added `FluxKontextBlock` in `backend/blocks/flux_kontext.py`
* Uses `ReplicateClient` to call Flux Kontext Pro or Max models
* Supports inputs for `prompt`, `input_image`, `aspect_ratio`, `seed`,
and `model`
* Outputs transformed image URL or error
* Added credit pricing logic for Flux Kontext models to
`block_cost_config.py`:
* Pro: 10 credits
* Max: 20 credits
* Added documentation for the new block at
`docs/content/platform/blocks/flux_kontext.md`
* Updated block index at `docs/content/platform/blocks/blocks.md` to
include Flux Kontext
---
### Checklist 📋
#### For code changes:
* [x] I have clearly listed my changes in the PR description
* [x] I have made a test plan
* [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
* [x] Prompt-only input generates an image
* [x] Prompt with image applies edit correctly
* [x] Image respects specified aspect ratio
* [x] Invalid image URL returns helpful error
* [x] Using the same seed gives consistent output
* [x] Output chaining works: result URI can be used in downstream blocks
* [x] Output from Max model shows higher fidelity than Pro
<details>
<summary>Example test plan</summary>
* [x] Create from scratch and execute an agent using Flux Kontext with
at least 3 blocks
* [x] Import agent with Flux Kontext from file upload, and confirm
execution
* [x] Upload agent (with Flux Kontext block) to marketplace (internal
test)
* [x] Import agent from marketplace and confirm correct execution
* [x] Edit agent with Flux Kontext block from monitor and confirm output
</details>
#### For configuration changes:
* [x] `.env.example` is updated or already compatible with my changes
* [x] `docker-compose.yml` is updated or already compatible with my
changes
* [x] I have included a list of my configuration changes in the PR
description (under **Changes**)
* No new environment variables or services introduced
<details>
<summary>Examples of configuration changes</summary>
* N/A
</details>
---------
Co-authored-by: Zamil Majdy <zamil.majdy@agpt.co>
### Changes 🏗️
This PR adds `Run 10 agents` step to wallet tasks that can be done by
running any agents 10 times either from Library or Builder (onboarding
agent run also counts).
- Merge `Finish onboarding` and `See results` steps into one in the
wallet
- User is redirected directly to onboarding agent runs in Library after
congrats screen
- Add `RUN_AGENTS` step and `agentRuns` integer to schema and related
migration
- Running agent from Library and Builder increments `agentRuns`
- Open NPS survey popup when 10 agents are run
- Fix resuming onboarding on login when unfinished
- Remove no longer needed `get-results.mp4` tutorial video
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Onboarding can be completed and proper reward is awarded
- [x] `Run 10 agents` can be completed and reward is awarded
- [x] When unning different agents and the same agent
- [x] Running from library and builder counts
- [x] Onboarding is resumed to last finished step on login
This makes button on the marketplace listing page show `See runs` if
user has an agent in the library.
### Changes 🏗️
- Remove `/` from the related endpoint
- Use `active_version_id` instead of `store_listing_version_id` to check
for the library agent
- Fix `get_store_agent_details`
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- Log in and pick an agent that has never been in user library
- [x] Button says `Add to library`
- Add the agent and return to the listing page
- [x] Button says `See runs`
- Remove agent from library
- [x] Button says `Add to library`
- Add agent again
- [x] Button says `See runs`
---------
Co-authored-by: Zamil Majdy <zamil.majdy@agpt.co>
- Resolves#9752
- Follow-up fix to #9940
### Changes 🏗️
- `GRAPH_EXECUTION_INCLUDE` ->
`graph_execution_include(include_block_ids)`
- Add `get_io_block_ids()` and `get_webhook_block_ids()` to
`backend.data.blocks`
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [ ] I have tested my changes according to the test plan:
- [ ] Payload for webhook-triggered runs is shown on
`/library/agents/[id]`
# Query Optimization for AgentNodeExecution Tables
## Overview
This PR describes the database index optimizations applied to improve
the performance of slow queries in the AutoGPT platform backend.
## Problem Analysis
The following queries were identified as consuming significant database
resources:
### 1. Complex Filtering Query (19.3% of total time)
```sql
SELECT ... FROM "AgentNodeExecution"
WHERE "agentNodeId" = $1
AND "agentGraphExecutionId" = $2
AND "executionStatus" = $3
AND "id" NOT IN (
SELECT "referencedByInputExecId"
FROM "AgentNodeExecutionInputOutput"
WHERE "name" = $4 AND "referencedByInputExecId" IS NOT NULL
)
ORDER BY "addedTime" ASC
```
### 2. Multi-table JOIN Query (8.9% of total time)
```sql
SELECT ... FROM "AgentNodeExecution"
LEFT JOIN "AgentNode" ON ...
LEFT JOIN "AgentBlock" ON ...
WHERE "AgentBlock"."id" IN (...)
AND "executionStatus" != $11
AND "agentGraphExecutionId" IN (...)
ORDER BY "queuedTime" DESC, "addedTime" DESC
```
### 3. Bulk Graph Execution Queries (multiple variations, ~10% combined)
Multiple queries filtering by `agentGraphExecutionId` with various
ordering requirements.
## Optimization Strategy
### 1. Composite Indexes for AgentNodeExecution
Set the following composite indexes to the `AgentNodeExecution` model:
```prisma
@@index([agentGraphExecutionId, agentNodeId, executionStatus])
@@index([addedTime, queuedTime])
```
#### Benefits:
- **Index 1**: Covers the exact WHERE clause of the complex filtering
query, allowing index-only scans
- **Index 2**: Optimizes queries filtering by graph execution and status
- **Index 3**: Supports efficient sorting when filtering by graph
execution
- **Index 4**: Optimizes ORDER BY operations on time fields
### 2. Composite Index for AgentNodeExecutionInputOutput
Added the following composite index:
```prisma
// Input and Output pin names are unique for each AgentNodeExecution.
@@unique([referencedByInputExecId, referencedByOutputExecId, name])
@@index([referencedByOutputExecId])
// Composite index for `upsert_execution_input`.
@@index([name, time])
```
#### Benefits:
- Dramatically improves the NOT IN subquery performance in Query 1
- Allows the database to use an index scan instead of a full table scan
- Reduces the subquery execution time from O(n) to O(log n)
## Expected Performance Improvements
1. **Query 1 (19.3% of total time)**:
- Expected improvement: 80-90% reduction in execution time
- The composite index on `[agentNodeId, agentGraphExecutionId,
executionStatus]` will allow index-only scans
- The subquery will benefit from the new index on
`AgentNodeExecutionInputOutput`
2. **Query 2 (8.9% of total time)**:
- Expected improvement: 50-70% reduction in execution time
- The `[agentGraphExecutionId, executionStatus]` index will reduce the
initial filtering cost
3. **Bulk Queries (10% combined)**:
- Expected improvement: 60-80% reduction in execution time
- Composite indexes including time fields will optimize sorting
operations
## Migration Considerations
1. **Index Creation Time**: Creating these indexes on existing large
tables may take time
2. **Storage Impact**: Each index requires additional storage space
3. **Write Performance**: Slight decrease in INSERT/UPDATE performance
due to index maintenance
## Additional Optimizations
### NotificationEvent Table Index
Added index for notification batch queries:
```prisma
@@index([userNotificationBatchId])
```
This optimizes the query:
```sql
SELECT ... FROM "NotificationEvent"
WHERE "userNotificationBatchId" IN (...)
```
#### Benefits:
- Eliminates full table scans when filtering by batch ID
- Improves query performance from O(n) to O(log n)
- Particularly beneficial for users with many notification events
## Future Optimizations
Consider these additional optimizations if needed:
1. Partitioning `AgentNodeExecution` table by `createdAt` or
`agentGraphExecutionId`
2. Implementing materialized views for frequently accessed aggregate
data
3. Adding covering indexes for specific query patterns
4. Implementing query result caching at the application level
---------
Co-authored-by: Zamil Majdy <zamil.majdy@agpt.co>
- Resolves#10024
Caching the repeated DB calls by the graph lifecycle hooks significantly
speeds up graph update/create calls with many authenticated blocks
(~300ms saved per authenticated block)
### Changes 🏗️
- Add and use `IntegrationCredentialsManager.cached_getter(user_id)` in
lifecycle hooks
- Split `refresh_if_needed(..)` method out of
`IntegrationCredentialsManager.get(..)`
- Simplify interface of lifecycle hooks: change `get_credentials`
parameter to `user_id`
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Save a graph with nodes with credentials
This pull request refines the handling of `input_data.content` and
improves error message formatting in the `run` method of `mem0.py`. The
changes enhance robustness and clarity in the code.
### Handling `input_data.content`:
* Updated the `run` method to handle `Content` objects explicitly,
ensuring proper formatting of messages when `input_data.content` is of
type `Content`. Additionally, non-standard types are now converted to
strings for consistent handling.
(`[autogpt_platform/backend/backend/blocks/mem0.pyR127-R130](diffhunk://#diff-d7abf8c3299388129480b6a9be78438fe7e0fbe239da630ebb486ad99c80dd24R127-R130)`)
### Error message formatting:
* Simplified the error message formatting by removing the unnecessary
`object=` keyword in the `str()` conversion of exceptions.
(`[autogpt_platform/backend/backend/blocks/mem0.pyL155-R157](diffhunk://#diff-d7abf8c3299388129480b6a9be78438fe7e0fbe239da630ebb486ad99c80dd24L155-R157)`)
## Summary
- fix AddMemoryBlock so `Content` input uses the underlying string
- improve error handling in Mem0 AddMemoryBlock
## Testing
- `ruff check autogpt_platform/backend/backend/blocks/mem0.py`
- `pre-commit run --files
autogpt_platform/backend/backend/blocks/mem0.py` *(fails: unable to
fetch remote hooks)*
- `poetry run pytest -k AddMemoryBlock -q` *(fails: Error 111 connecting
to localhost:6379)*
Checklist 📋
For code changes:
I have clearly listed my changes in the PR description
I have made a test plan
I have tested my changes according to the test plan:
Payload for webhook-triggered runs is shown on /library/agents/[id]
### Changes 🏗️
Keep the original URL when an HTTP error occurs in
`SendWebRequestBlock`.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Test sending POST request on a web that doesn't support POST
request using `SendWebRequestBlock`.
The executor can sometimes become dangling due to the executor stopping
executing messages but the process is not fully killed. This PR avoids
such a scenario by simply keeping retrying it.
### Changes 🏗️
Introduced continuous_retry decorator and use it to executor message
consumption/
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] Run executor service and execute some agents.
Makes all optional fields on `Credentials` models actually optional, and
sets `exclude_none=True` on the corresponding `model_dump`.
This is a hotfix: after running the `aryshare-revid` branch on the dev
deployment, there is some data in the DB that isn't valid for the
`UserIntegrations` model on the `dev` branch (see
[here](https://github.com/Significant-Gravitas/AutoGPT/pull/9946#discussion_r2098428575)).
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] This fix worked on the `aryshare-revid` branch:
52b6d9696b
- Resolves#9987
### Changes 🏗️
- Split `pin_url(..)` out of `validate_url(..)` and call
`extra_url_validator` in between
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] GitHub Read Pull Request Block works with "Include PR Changes"
enabled
Resolves#9947
### Changes 🏗️
Backend:
- Send a graph execution update after terminating a run
- Don't wipe the graph execution stats when not passed in to `update_graph_execution_stats`
Frontend:
- Don't hide the output of stopped runs
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- Go to `/library/agents/[id]`
- Run an agent that takes a while (long enough to click stop and see the effect)
- Hit stop after it has executed a few nodes
- [x] -> run status should change to "Stopped"
- [x] -> run stats (steps, duration, cost) should stay the same or increase only one last time
- [x] -> output so far should be visible
- [x] -> shown information should stay the same after refreshing the page
---
Co-authored-by: Krzysztof Czerwinski <34861343+kcze@users.noreply.github.com>
