deco Studio

Monitoring

See every tool call your agents make — who, what, when, and how it performed

Every tool call, in one place

Studio logs every tool invocation that flows through it — automatically, with no setup required. When an agent queries GitHub, sends a Slack message, or calls your database, there’s a record of it.

You’ll find it under Settings → Monitor. Every agent’s tool calls roll up there, organization-wide.

What gets logged: Tool calls that go through your connections and agents (GitHub queries, database operations, Shopify calls, etc.). Studio management actions — like creating a connection or updating settings — are not included in the monitoring log.

Without a central place for this, you’d be correlating logs from different clients and different services, each in a different format — if they were logging at all. Studio captures everything at the point where all traffic passes through, so you always have a complete, consistent picture.

What you can see

For every tool call, Studio records:

Who made it

  • The user or API key that made the request
  • Which agent or automation the call came from

What happened

  • The tool that was called
  • The inputs that were sent
  • The output that came back
  • Whether it succeeded or failed — and why, if it didn’t

When and how long

  • Timestamp
  • How long the call took, in milliseconds

What you can do with it

Debug failures fast

When something breaks, open Settings → Monitor and see exactly what happened — what inputs were sent, what error came back, when it occurred. No SSH, no log files, no guessing.

Track performance

See which tools are slowest (p50 / p95 latency), which are called most often, and where time is being spent. This makes it easy to spot bottlenecks before they become problems.

Understand how your team uses agents

See which tools are actually getting used versus which are just configured. Spot where agents are retrying or hitting errors frequently. Use this to improve your agents over time.

Audit access

Every tool call is tied to a specific user or API key. If you need to know who accessed what and when — for compliance, security review, or incident response — it’s all there.

A note on sensitive data

Tool inputs and outputs are logged by default. If you’re working with sensitive data (customer records, financial data, etc.), use scoped API keys to limit what each agent can access, and review your organization’s data retention settings.

Studio provides the visibility — your team decides the policies.

Found an error or want to improve this page?

Edit this page