Agent Observability Platform

See what your AI agents are actually doing.

Open-source, self-hosted observability, risk signals, and behavior governance for autonomous AI agents. Guard what they do before they do it.

Up and running in 5 minutes

Three steps from install to full agent observability.

1

Install the SDK

npm install dashclaw

Zero dependencies. Works with any Node.js agent.

2

Initialize your agent

const claw = new DashClaw({
  apiKey: '...',
  agentId: 'my-agent',
})

One constructor. Your API key scopes all data.

3

See it on the dashboard

await claw.createAction({
  action_type: 'deploy',
  declared_goal: 'Ship auth',
})

Actions, signals, and loops appear instantly.

Everything you need to trust your agents

Built for teams running autonomous AI agents in production.

Action Recording & Audit Trail

Every agent action logged with type, goal, risk score, status, outcome, and systems touched. Full post-mortem with root-cause tracing.

7 Risk Signals

Automatic detection of autonomy spikes, repeated failures, stale loops, assumption drift, and more. No configuration required.

Agent Workspace

Session handoffs, context threads, key points, automation snippets, user preferences, and daily digests in one tabbed interface.

Behavior Guard

Agents check in before risky actions. Set risk thresholds, require approvals, and block dangerous operations — without changing agent code.

Session Handoffs

Structured handoff documents for continuity between agent sessions.

Context Manager

Capture key points and organize context into threads for long-running topics.

Open Loop Tracking

Track unresolved dependencies, pending approvals, and blockers across agents.

Assumption Monitoring

Log what agents assume, validate or invalidate, and catch drift early.

Security Scanning

18 regex patterns detect API keys, tokens, and PII before data leaves your system.

Memory Health

Track memory file counts, duplicates, stale facts, entities, and topics over time.

Daily Digest

Aggregated daily summary from actions, decisions, lessons, content, and goals.

57-Method SDK

Zero-dependency npm package. ESM + CJS. Drop into any Node.js agent in minutes.

57 methods across 13 categories

One SDK. Full observability.

Install from npm. Zero dependencies. Works with any Node.js agent framework. Actions, handoffs, context, snippets, messaging, security scanning, and more.

npm packageNode.jsESM + CJSZero Dependencies
View full SDK docs
// instrument your agent
import { DashClaw } from 'dashclaw'
const claw = new DashClaw({
apiKey: process.env.DASHCLAW_API_KEY,
agentId: 'my-agent',
})
// check guard before acting
const { decision } = await claw.guard({
action_type: 'deploy',
risk_score: 85,
})
// record an action
await claw.createAction({
action_type: 'deploy',
declared_goal: 'Ship auth service',
})
// create a session handoff
await claw.createHandoff({
summary: 'Completed auth system',
key_decisions: ['JWT over sessions'],
})
// bulk sync agent state
await claw.syncState({ goals, learning, snippets })

7 built-in risk signals

Automatic detection of problematic agent behavior. No configuration required.

Autonomy Spike

SIGNAL-01

Agent taking too many actions without human checkpoints

High Impact, Low Oversight

SIGNAL-02

Critical actions without sufficient review

Repeated Failures

SIGNAL-03

Same action type failing multiple times

Stale Loop

SIGNAL-04

Open loops unresolved past their expected timeline

Assumption Drift

SIGNAL-05

Assumptions becoming stale or contradicted by outcomes

Stale Assumption

SIGNAL-06

Assumptions not validated within expected timeframe

Stale Running Action

SIGNAL-07

Actions stuck in running state for over 4 hours

Production-ready operations

Team management, audit trails, webhooks, and more — built in from day one.

Team Management

Invite links, role-based access (admin/member), and workspace isolation.

Webhooks & Alerts

HMAC-signed webhook delivery plus email alerts via Resend for signal notifications.

Activity Audit Log

Every admin action logged — key creation, invites, role changes, and usage activity.

Guided Onboarding

4-step checklist: create workspace, generate key, install SDK, send first action.

Multi-Tenant

Full org isolation with API key scoping, per-agent settings, and org management.

Agent Tools

20+ Python CLI tools for local ops with optional --push sync to the dashboard.

20+ Python CLI tools

Local Agent Toolkit

Python CLI tools that run alongside your agent. Local-first with SQLite storage. Add --push to sync anything to your dashboard.

Learning & Decisions

Log decisions, lessons, and outcomes. Track what worked and why.

learner.py log "Used JWT" --push

Context & Handoffs

Key points, threads, and session continuity documents.

context.py capture "Dark theme" --push

Memory & Health

Scan memory files, track entities, detect stale facts.

scanner.py scan ~/.agent/memory --push

Goals & Relationships

Goal milestones, contacts, interactions, and follow-ups.

goals.py add "Ship auth" --push

Security & Audit

Outbound content filtering, session isolation, audit logging.

outbound_filter.py scan message.txt --push

Automation & Snippets

Reusable code snippets with search, tags, and use tracking.

snippets.py add "retry logic" --push

Start monitoring in 5 minutes

Install the SDK, send your first action, and see signals on the dashboard. Open-source and self-hosted.