Compatibility¶
Compatibility reference
Check here for compatibility status. Provider and framework pages have usage details but may not reflect the latest support status.
Providers¶
| Provider | Status | Docs |
|---|---|---|
| OpenAI | Supported | OpenAI |
| Anthropic | Supported | Anthropic |
| Gemini | Supported | Gemini |
HTTP libraries¶
Tenro intercepts HTTP requests made through httpx only. Most LLM provider SDKs (OpenAI, Anthropic, Google, Mistral) use httpx by default, so they work automatically.
| HTTP library | Status | Notes |
|---|---|---|
| httpx | Supported | Used by OpenAI, Anthropic, Google, Mistral SDKs |
| requests | Not supported | Use @link_llm with target= as workaround |
| aiohttp | Not supported | Use @link_llm with target= as workaround |
| Custom HTTP clients | Not supported | Use @link_llm with target= as workaround |
Endpoints¶
Simulation is supported only for endpoint families explicitly listed below. Anything not listed is unsupported.
| Endpoint family | Status |
|---|---|
| Text generation (chat-style requests) | Supported |
| Embeddings | Not supported |
| Images | Not supported |
| Audio / TTS / STT | Not supported |
| Realtime / WebSockets | Not supported |
Covered request shapes include chat-style prompts/messages (e.g., OpenAI chat.completions, Anthropic messages). Other API families (assistants, responses, batch) are unsupported unless listed here.
Feature coverage (streaming, tool calling, etc.) within supported endpoints is provider-specific. See provider pages for details.
Frameworks¶
| Framework | Status | Notes |
|---|---|---|
| LangChain | Experimental | Works when using OpenAI/Anthropic with httpx |
| CrewAI | Experimental | Works when using OpenAI/Anthropic with httpx |
| LangGraph | Experimental | Works when using OpenAI/Anthropic with httpx |
| Pydantic AI | Experimental | Works when using OpenAI/Anthropic with httpx |
| AutoGen | Experimental | Works when using OpenAI/Anthropic with httpx |
| LlamaIndex | Experimental | Works when using OpenAI/Anthropic with httpx |
Support definitions¶
- Supported: Tested before every release. We fix breakages promptly.
- Experimental: Tested, but may break when frameworks update.
Version policy¶
Tenro is pre-1.0. The API may change between releases. Breaking changes are noted in release notes.
When interception doesn't apply¶
If your simulated response isn't returned:
- Check you're using a supported provider (see table above)
- Check the SDK uses httpx transport (default for OpenAI/Anthropic)
- Or use
@link_llmwithtarget=to route simulations directly
See Troubleshooting for more help.
See also¶
- How Tenro works - HTTP interception explained
- LLM Providers - Provider-specific details
- Frameworks - Framework-specific examples