Capability Matrix
This page renders the public provider operation matrix. It shows customer-facing operation support by provider, mode, modality, and streaming availability.
The matrix is intentionally limited to integration planning details. It does not document adapter internals, routing heuristics, detector behavior, or provider-key handling.
| Provider | Operation | Modes | Modalities | Streaming |
|---|---|---|---|---|
| Anthropic | code_execution | async, sync | text | No |
| Anthropic | computer.use | realtime | computer | No |
| Anthropic | generate.text | async, stream, sync | image, text | Yes |
| Anthropic | understand.image | async, stream, sync | image | Yes |
embed.text | async, sync | text | No | |
generate.text | async, stream, sync | image, text | Yes | |
google_gemini_tool_run_governance | sync | text | No | |
run.batch | async | text | No | |
transcribe.audio | async, sync | audio | No | |
understand.image | async, stream, sync | image | Yes | |
| MCP | mcp_tool_call_evidence_emission | sync | text | No |
| OpenAI | code_execution | async, sync | text | No |
| OpenAI | embed.text | async, sync | text | No |
| OpenAI | generate.audio | async, sync | audio | No |
| OpenAI | generate.image | async, sync | image | No |
| OpenAI | generate.text | async, stream, sync | image, text | Yes |
| OpenAI | openai_responses_tool_run_governance | sync | text | No |
| OpenAI | run.batch | async | text | No |
| OpenAI | transcribe.audio | async, sync | audio | No |
| OpenAI | understand.image | async, stream, sync | image | Yes |
| Search | search_policy_evidence | sync | text | No |
| xAI | generate.text | async, stream, sync | text | Yes |
| xAI | run.batch | async | text | No |
Use Provider Support for provider-level caveats and Executions for request examples.
Last updated on Edit this page on GitHub