chat
Read-only
Send messages to an LLM with automatic output verification across 6 dimensions (structure, latency, cost, etc.) and auto-healing on failure.
Instructions
Send a chat message to an LLM with automatic output verification. Routes through the best available provider, validates the response across 6 dimensions (structure, schema, latency, cost, identity, integrity), and auto-heals on failure by retrying or failing over to another provider. Returns the response text plus a validation report showing which dimensions passed or failed.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| model | No | Model name or 'auto' for automatic provider selection. Examples: 'gpt-4o-mini', 'claude-3-haiku-20240307', 'deepseek-chat'. Default: 'auto'. | |
| messages | Yes | Conversation messages in OpenAI format: [{role: 'user', content: '...'}, ...]. Each message must have 'role' (system/user/assistant) and 'content' (string). | |
| provider | No | Force a specific provider: 'openai', 'anthropic', 'deepseek', 'moonshot', 'zhipu', 'qwen', 'siliconflow', 'groq', 'together'. If omitted, auto-selects by priority and health. | |
| max_tokens | No | Maximum tokens in response. Limits output length to control cost and latency. | |
| temperature | No | Sampling temperature (0.0-2.0). Lower values for more deterministic output. Default: provider-specific. | |
| system_prompt | No | System prompt to prepend to the conversation. Useful for setting context, role, or output format requirements. |