API reference

Chat Completions

Generate text with the OpenAI-compatible chat interface.

POST/v1/chat/completionsapi.17.wtf

On this page

Request body Messages & content Example request Read the response Compatibility

Request body

Parameter	Type / requirement	Description
`model`	string · required	Exact text model ID from the catalog.
`messages`	array · required	Conversation messages with role and content.
`max_tokens`	integer · optional	Positive output limit; defaults to 4096.
`max_completion_tokens`	integer · optional	Alternative output limit; takes precedence over max_tokens.
`temperature`	number · optional	Sampling control; accepted range depends on model.
`top_p`	number · optional	Nucleus sampling; model dependent.
`stop`	string or string[] · optional	Stop sequences.
`stream`	boolean · optional	Defaults to false. Use true for SSE.
`tools`	array · optional	Function definitions with name, description, parameters.
`tool_choice`	string or object · optional	auto, required, none, or a named function.
`reasoning_effort`	string · optional	Model-specific effort; reasoning.effort is an alias.

Supported roles are system, developer, user, assistant, and tool. System and developer messages become the system prompt. Content can be text or an array of text and image_url parts; vision requires a capable model. Image URLs can be HTTPS URLs or base64 data URLs. Assistant tool calls use tool_calls; tool results use the tool role and matching tool_call_id.

Example request

Replace MODEL_ID with a current text model ID.

curl https://api.17.wtf/v1/chat/completions \
  -H "Authorization: Bearer $LM0_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "MODEL_ID",
  "messages": [
    {
      "role": "user",
      "content": "Hello!"
    }
  ],
  "max_tokens": 512
}'

Read the response

The non-streaming response contains id, object, created, model, choices, and usage. Read text from choices[0].message.content; function calls are in choices[0].message.tool_calls. Check finish_reason before treating a generation as complete. Usage includes prompt_tokens, completion_tokens, and total_tokens. Reasoning-capable models may include reasoning_content.

Compatibility

LM0 translates the documented fields into a shared request format. Additional provider-specific fields are not guaranteed to be forwarded. Multiple choices, audio, embeddings, and arbitrary provider extensions are not documented capabilities of this endpoint. Test optional parameters against your chosen model.