ChemAgent

MODEL PROVIDER

GLM

Text and vision models with speech recognition and embedding support.

Provider profile

AreaCurrent documentationOperational meaning
Representative modelsglm-5.2 / glm-5v-turboChoose the exact model available to the authorized account.
Capability scopeText, Vision, Speech-to-text, EmbeddingEach capability is configured independently in the Web console.
ConnectionProvider adapter or approved compatible endpointCredentials and base URLs stay inside protected deployment configuration.

Supported capabilities

  • Text
  • Vision
  • Speech-to-text
  • Embedding
Independent capability routing

Chat, vision, image, speech, embedding, and search may use different providers and model identifiers; selecting this provider for one capability does not force it onto the others.

Configure in the Web console

  1. 1
    Open Models

    Select the provider and review the capability cards it can supply.

  2. 2
    Add connection details

    Enter the authorized credential and service address without placing secrets in chat or project files.

  3. 3
    Choose models by capability

    Set the chat model first, then configure vision, image, speech, embedding, or search only where supported.

  4. 4
    Save and test

    Run a bounded test for authentication, streaming, tool calls, cancellation, and usage reporting before production use.

Production checklist

  • ✓
    Availability

    Confirm the exact model identifier and account entitlement in the target region.

  • ✓
    Behaviour

    Verify tool-call schemas, structured output, context limits, and error handling.

  • ✓
    Governance

    Review retention, data location, access control, latency, and budget.

  • ✓
    Observability

    Confirm whether the integration returns token-usage fields used by ChemAgent statistics.