INTEGRATIONS / PROPOSED SUPPORT

Start at the source of the action.

We’re exploring runtime capture and assistant-reported workflows. Integration access begins with a scoped discussion.

No runtime connector is publicly available for UpTrain. The options below are proposed directions, not tested compatibility claims. Third-party names do not imply partnership.

RUNTIME CAPTURE · PROPOSED

Hermes Agent

Evaluate runtime lifecycle events, tool activity, failures, and recovery within an explicitly scoped pilot.

Explore the approach RUNTIME CAPTURE · PROPOSED

OpenAI Agents SDK

Explore instrumentation for TypeScript agent workflows and the boundaries of emitted events.

Explore the approach COMPATIBILITY DISCUSSION · PROPOSED

Coding-agent runtimes

Discuss your runtime and version, including Codex CLI or Claude Code. Support must be verified rather than assumed.

Explore the approach AGENT-REPORTED SESSIONS · PROPOSED

ChatGPT / MCP

Explore checkpoint reporting with a separate evidence class and explicit tracking consent.

Explore the approach

Evidence and machine interfaces

Our researched scope includes canonical evidence bundles, optional independent attestation, REST access, and scoped MCP interfaces. NexArt integration is under consideration; no provider connection or certification guarantee is active.

Before a pilot, we need to agree capture coverage, exclusions, credentials, request limits, retention, and permissions. Read access must not automatically grant write or spending authority.

Read the developer preparation guide →

LET’S START WITH YOUR WORKFLOW

Make your next agent run easier to understand.

Discuss early access

Tell us what your agents do and what you need to understand. No payment or runtime connection required.