> ## Documentation Index
> Fetch the complete documentation index at: https://developers.namuai-lab.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Converse

> One turn of a spoken conversation in Hausa: send what the person said, get back what was heard, the reply as text and the reply spoken. A research preview. New keys do not include the `converse` scope: tick it when you create the key. Counts the audio sent plus the reply.



## OpenAPI

````yaml /api-reference/openapi.yaml post /v1/converse
openapi: 3.1.0
info:
  title: Namu API
  version: 1.0.0
  description: >-
    Speech recognition, speech synthesis, translation and a voice agent for
    Hausa, Zarma and French. Describes the API as it behaves on 8 November 2026.
  contact:
    name: Namu AI Lab
    email: contact@namuai.org
    url: https://www.namuai-lab.com
servers:
  - url: https://api.namuai-lab.com
    description: Production
security:
  - bearerAuth: []
tags:
  - name: Speech
    description: Text to speech.
  - name: Transcription
    description: Speech to text.
  - name: Translation
    description: Hausa and French, from text or speech.
  - name: Voice agent
    description: A spoken conversation in Hausa. Research preview.
  - name: Models and voices
    description: What the API serves. No key needed.
  - name: Account
    description: Your usage this month.
paths:
  /v1/converse:
    post:
      tags:
        - Voice agent
      summary: Converse
      description: >-
        One turn of a spoken conversation in Hausa: send what the person said,
        get back what was heard, the reply as text and the reply spoken. A
        research preview. New keys do not include the `converse` scope: tick it
        when you create the key. Counts the audio sent plus the reply.
      operationId: createConversationTurn
      requestBody:
        required: true
        content:
          multipart/form-data:
            schema:
              type: object
              required:
                - file
              properties:
                file:
                  type: string
                  format: binary
                  description: What the person said.
                persona:
                  type: string
                  enum:
                    - nadia
                  default: nadia
                  description: Who answers. `nadia` is the only persona.
                session_id:
                  type: string
                  maxLength: 100
                  description: Your own id for the conversation. Echoed back, never stored.
      responses:
        '200':
          description: The turn.
          headers:
            X-Request-Id:
              $ref: '#/components/headers/RequestId'
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ConverseTurn'
        '400':
          $ref: '#/components/responses/BadRequest'
        '401':
          $ref: '#/components/responses/Unauthorized'
        '403':
          $ref: '#/components/responses/Forbidden'
        '413':
          $ref: '#/components/responses/TooLarge'
        '415':
          $ref: '#/components/responses/UnsupportedAudio'
        '422':
          $ref: '#/components/responses/NothingHeard'
        '429':
          $ref: '#/components/responses/TooManyRequests'
        '502':
          $ref: '#/components/responses/Upstream'
components:
  headers:
    RequestId:
      description: A unique id for this request. Include it when you write to us.
      schema:
        type: string
        example: 3f2a9c1e5b7d4e8fa0c2b6d9e1f3a5c7
  schemas:
    ConverseTurn:
      type: object
      properties:
        model:
          type: string
          example: voice-agent
        persona:
          type: string
          example: nadia
        session_id:
          type:
            - string
            - 'null'
        heard:
          type: string
          description: What the person said, as text.
        reply:
          type: string
          description: The answer, as text.
        audio:
          type: string
          description: The answer spoken, base64 WAV.
        sample_rate:
          type: integer
          example: 24000
        audio_seconds:
          type: number
          description: Audio sent plus the reply.
    Error:
      type: object
      properties:
        error:
          type: object
          properties:
            type:
              type: string
              description: A stable code, such as `invalid_api_key`.
            message:
              type: string
              description: What happened, in plain words.
            request_id:
              type: string
  responses:
    BadRequest:
      description: >-
        The request was not usable: `invalid_request`, `invalid_json`,
        `missing_text`, `missing_file`, `empty_audio`, `unknown_voice`,
        `unsupported_language`, `unsupported_pair`, `unknown_persona`.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
          example:
            error:
              type: unknown_voice
              message: 'Unknown voice. Try one of: nadia, warm, fr, mouhamad.'
              request_id: 3f2a9c1e5b7d4e8fa0c2b6d9e1f3a5c7
    Unauthorized:
      description: >-
        No key, not a Namu key, or a revoked key: `missing_api_key`,
        `invalid_api_key`, `key_revoked`.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
          example:
            error:
              type: missing_api_key
              message: 'Send your key as ''Authorization: Bearer namu_sk_...''.'
              request_id: 3f2a9c1e5b7d4e8fa0c2b6d9e1f3a5c7
    Forbidden:
      description: >-
        `insufficient_scope` (the key may not call this endpoint) or
        `account_paused`.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    TooLarge:
      description: >-
        `payload_too_large` (over 25 MB), `audio_too_long` (over 300 s) or
        `text_too_long` (over 2,000 characters).
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    UnsupportedAudio:
      description: >-
        `unsupported_audio`: the file could not be read. Send WAV, MP3, M4A,
        AAC, OGG, WebM or FLAC.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    NothingHeard:
      description: '`nothing_heard`: no speech was found in the audio.'
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    TooManyRequests:
      description: >-
        `rate_limited` (too many requests this minute; see `Retry-After`) or
        `quota_exceeded` (this month's allowance is used).
      headers:
        Retry-After:
          description: Seconds to wait. Sent with `rate_limited` only.
          schema:
            type: integer
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    Upstream:
      description: >-
        `upstream_error`: the model did not answer. Try again; a request that
        fails counts nothing.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      description: >-
        Your secret key: `Authorization: Bearer namu_sk_...`. Create one at
        https://www.namuai-lab.com/account/keys.

````

This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.