mirror of https://github.com/danielmiessler/Fabric.git synced 2026-02-19 10:14:21 -05:00

Files

Kayvan Sylvan 0894ec28c1 chore: update model references to latest versions across docs and code

- Update default summarize model to `claude-sonnet-4-5`
- Replace `gpt-4o` references with `gpt-5.2` throughout docs
- Replace `gpt-4o-mini` references with `gpt-5-mini`
- Add MiniMax-M2.5 and M2.5-lightning to static models list
- Update image generation warning to suggest `gpt-5.2`
- Update OpenAI chat example script to use `gpt-5-mini`
- Update REST API docs and examples with current model names

2026-02-12 16:45:05 -08:00

2.5 KiB

Raw Blame History

REST API Pattern Variables Example

This example demonstrates how to use pattern variables in REST API calls to the /chat endpoint.

Example: Using the `translate` pattern with variables

Request

{
  "prompts": [
    {
      "userInput": "Hello my name is Kayvan",
      "patternName": "translate",
      "model": "gpt-5.2",
      "vendor": "openai",
      "contextName": "",
      "strategyName": "",
      "variables": {
        "lang_code": "fr"
      }
    }
  ],
  "language": "en",
  "temperature": 0.7,
  "topP": 0.9,
  "frequencyPenalty": 0.0,
  "presencePenalty": 0.0
}

Pattern Content

The translate pattern contains:

You are an expert translator... translate them as accurately and perfectly as possible into the language specified by its language code {{lang_code}}...

...

- Translate the document as accurately as possible keeping a 1:1 copy of the original text translated to {{lang_code}}.

{{input}}

How it works

The pattern is loaded from patterns/translate/system.md
The {{lang_code}} variable is replaced with "fr" from the variables map
The {{input}} placeholder is replaced with "Hello my name is Kayvan"
The resulting processed pattern is sent to the AI model

Expected Result

The AI would receive a prompt asking it to translate "Hello my name is Kayvan" to French (fr), and would respond with something like "Bonjour, je m'appelle Kayvan".

Testing with curl

curl -X POST http://localhost:8080/api/chat \
  -H "Content-Type: application/json" \
  -d '{
    "prompts": [
      {
        "userInput": "Hello my name is Kayvan",
        "patternName": "translate",
        "model": "gpt-5.2",
        "vendor": "openai",
        "variables": {
          "lang_code": "fr"
        }
      }
    ],
    "temperature": 0.7
  }'

Multiple Variables Example

For patterns that use multiple variables:

{
  "prompts": [
    {
      "userInput": "Analyze this business model",
      "patternName": "custom_analysis",
      "model": "gpt-5.2",
      "variables": {
        "role": "expert consultant",
        "experience": "15",
        "focus_areas": "revenue, scalability, market fit",
        "output_format": "bullet points"
      }
    }
  ]
}

Implementation Details

Variables are passed in the variables field as a key-value map
Variables are processed using Go's template system
The {{input}} variable is automatically handled and should not be included in the variables map
Variables support the same features as CLI variables (plugins, extensions, etc.)

2.5 KiB Raw Blame History