Skip to content

Feature request: OpenAI-compatible server example (POST /v1/audio/transcriptions) #23

Description

@iswaryaalex

Hi @mudler and LocalAI team, this is a great project! Thank you!

Request:
OpenAI-API-compatible server binary, along the lines of whisper.cpp's whisper-server, perhaps living in examples/server/.

Scope
POST /v1/audio/transcriptions: multipart upload, response_format ∈ json | text | verbose_json. The existing --json output (text + per-word start/end/conf) maps almost 1:1 onto verbose_json.
GET /v1/models returning the loaded GGUF, so stock OpenAI SDKs work unmodified.

References

whisper.cpp server example: https://github.com/ggml-org/whisper.cpp/tree/master/examples/server
llama.cpp llama-server (OpenAI-compatible precedent in the ggml ecosystem): https://github.com/ggml-org/llama.cpp/tree/master/tools/server

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions