Reference
Unsupported features
Kyara Intelligence is text in, text out. These are the OpenAI features, endpoints and parameters it doesn't offer.
Kyara supports chat completions for text generation. Any OpenAI-compatible client can connect, but the capabilities below are not available.
Summary
- Function and tool calling:
toolsandtool_choiceare dropped, so the model never sees them. - Structured outputs and JSON mode:
response_formatis dropped. - Vision: image inputs are rejected. Message content must be plain text.
- Audio: no audio input or output, speech synthesis or transcription.
- Embeddings: there is no
/v1/embeddingsendpoint.
Available endpoints
These are the endpoints for API clients. Every other path returns 404 in the OpenAI error envelope.
| Endpoint | Status |
|---|---|
POST /v1/chat/completions | Available (text generation only) |
GET /v1/models | Available (no key needed) |
GET /v1/balance | Available (Kyara extension) |
/v1/completions (legacy) | Not available |
/v1/embeddings | Not available |
/v1/images/* | Not available |
/v1/audio/* | Not available |
/v1/responses | Not available |
/v1/files, /v1/batches | Not available |
/v1/fine_tuning/* | Not available |
/v1/moderations | Not available |
/v1/assistants, /v1/threads, /v1/vector_stores | Not available |
/v1/realtime | Not available |
Request parameters
Chat completions requests are checked against an allowlist. Every parameter is either passed to the model, dropped
without an error, or rejected with 400.
Supported
Accepted and passed to the model:
model,messagestemperature,top_p,top_k,min_p,presence_penalty,frequency_penalty,repetition_penalty,seedmax_tokens,max_completion_tokens,stop(up to four sequences)n(only1)stream,stream_options(include_usage,include_obfuscation)logprobs,top_logprobs,user,session_idreasoning(effort,max_tokens,enabled,exclude)reasoning_effort(folded intoreasoning.effort; an explicitreasoning.effortwins)
Dropped without an error
tools,tool_choice,functions,parallel_tool_callsresponse_format(JSON mode and structured outputs)logit_bias,predictionmodalities,audio,web_search_optionsstore,metadata,service_tier
Rejected with an error
These return a 400 validation error in the OpenAI envelope, with param naming the field; see Validation errors.
- Message
contentthat isn't a plain string, such as a content array withimage_urlor audio parts. - A message role other than
system,developer,userorassistant. There is notoolrole. - Unknown fields inside a message,
stream_optionsorreasoning. Messages accept onlyrole,content,name,reasoning,reasoning_contentandreasoning_details, so an assistant message carryingtool_callsis rejected. - An
nother than1, more than fourstopsequences, or asession_idlonger than 256 characters.
Notes
top_k,min_p,repetition_penaltyandsession_idare Kyara extensions beyond the OpenAI standard; see Chat completions.- Requests without
max_tokensormax_completion_tokensget an output limit of 8,192 tokens, and higher limits than the model allows are lowered to its maximum. - Support for sampling parameters varies by model.