Provider abstraction for OpenAI, Anthropic, and Ollama. Foundation for all AI plugins.

AI Core Plugin

The AI Core plugin provides a unified provider interface for LLM completions, embeddings, and streaming. It's the foundation that all other AI plugins depend on.

Installation

Copy the plugin directory into your FastCMS plugins/ folder:

cp -r ai_core/ plugins/ai_core/

Restart FastCMS. The plugin will be auto-detected.

Configuration

Configure the provider via the API or admin panel:

# Via API
curl -X POST http://localhost:8000/api/v1/plugins/ai/configure \
  -H "Content-Type: application/json" \
  -d '{
    "provider": "openai",
    "api_key": "sk-your-key-here",
    "model": "gpt-4o-mini"
  }'

Or set values in the plugin settings:

curl -X PATCH http://localhost:8000/api/v1/admin/plugins/ai-core/settings \
  -H "Content-Type: application/json" \
  -d '{
    "provider": "openai",
    "api_key": "sk-your-key-here",
    "model": "gpt-4o-mini"
  }'

Supported Providers

Provider	Completions	Embeddings	Streaming	Tool Calling
OpenAI	Yes	Yes	Yes	Yes
Anthropic	Yes	No (use separate embed provider)	Yes	Yes
Ollama	Yes	Yes	Yes	Yes

OpenAI

{
  "provider": "openai",
  "api_key": "sk-...",
  "model": "gpt-4o-mini"
}

Works with any OpenAI-compatible API (e.g., Azure OpenAI, local proxies) by setting base_url.

Anthropic

{
  "provider": "anthropic",
  "api_key": "sk-ant-...",
  "model": "claude-sonnet-4-20250514"
}

Anthropic doesn't provide an embedding API. Configure a separate embedding provider:

{
  "provider": "anthropic",
  "api_key": "sk-ant-...",
  "embed_provider": "openai",
  "embed_api_key": "sk-...",
  "embed_model": "text-embedding-3-small"
}

Ollama (Local)

{
  "provider": "ollama",
  "api_key": "",
  "base_url": "http://localhost:11434",
  "model": "llama3.1:8b"
}

No API key needed. Install Ollama and pull a model:

ollama pull llama3.1:8b
ollama pull nomic-embed-text  # for embeddings

API Endpoints

All endpoints are mounted at /api/v1/plugins/ai/.

GET /ai/status

Check if the provider is configured.

{
  "configured": true,
  "available_providers": ["openai", "anthropic", "ollama"],
  "provider": "openai",
  "model": "gpt-4o-mini"
}

POST /ai/generate

Generate a completion.

curl -X POST http://localhost:8000/api/v1/plugins/ai/generate \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "What is FastCMS?"}
    ],
    "temperature": 0.7,
    "max_tokens": 500
  }'

Response:

{
  "content": "FastCMS is an open-source Backend-as-a-Service...",
  "model": "gpt-4o-mini",
  "usage": {"prompt_tokens": 25, "completion_tokens": 50, "total_tokens": 75}
}

POST /ai/generate/stream

Stream a completion as Server-Sent Events.

curl -N http://localhost:8000/api/v1/plugins/ai/generate/stream \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [{"role": "user", "content": "Tell me a story"}]
  }'

POST /ai/configure

Configure the provider at runtime (no restart needed).

Using in Other Plugins

Other plugins can import from AI Core:

# In your plugin's route handler
from fastcms_plugin.ai_core.providers import get_provider, get_embed_provider

async def my_handler():
    provider = get_provider()
    result = await provider.complete([
        {"role": "user", "content": "Hello"}
    ])
    return result.content

Dependencies

None — uses httpx which is already included in FastCMS.

AI Core Plugin

On this page