Skip to content

Integrations

Connect the stack you already use

Bring your models, frameworks, data, and workflows. Block Convey fits around the AI app you already built.

Supported integration groups

Models and cloud AI

  • OpenAI

    Available

    Model calls, prompts, responses, and token usage

  • Anthropic

    Available

    Model calls, prompts, responses, and token usage

  • AWS Bedrock

    Available

    Hosted model activity and invocation metadata

  • Azure AI

    Available

    Hosted model activity and deployment metadata

Frameworks and tracing

  • LangChain

    Available

    Chains, tools, retrieval steps, and run traces

  • LangGraph

    Available

    Nodes, edges, state transitions, and agent runs

  • OpenTelemetry

    Available

    Any OTLP-compatible trace export

  • Custom SDK or API

    Available

    Your own instrumentation through the PRISM SDK or API

Data and ML

  • Databricks

    Coming soon

    Model and workflow activity from the lakehouse

  • Snowflake

    Coming soon

    Stored conversation and activity tables

  • MLflow

    Coming soon

    Tracked runs, models, and experiment metadata

  • Database and file imports

    Beta

    Supported database and file-based activity

Historical activity

  • OpenTelemetry exports

    Available

    Existing OTLP export files

  • JSON or CSV

    Available

    Exported conversations and runs

  • Supported tracing platforms

    Beta

    Trace history from a supported platform

  • Database conversations

    Beta

    Conversation tables you already store

Setup guides are published for each integration once the connection path, tested steps, and known limits are verified.

Do not see your stack?

Tell us what you use. We will confirm the best current path.