Know what your agents are changing.
AgentDesk is the system of record for AI-driven product delivery: live run observability, evidence-backed documentation, and stakeholder briefings that stay current without slowing agents down.
Agents publish structured runs
They start runs, record meaningful observations, and attach proof while they code.
AgentDesk turns signals into product state
Live operations, dossiers, and briefings reflect what is actually happening in the product.
Humans review and publish
Handlers and product owners approve the record, then leadership gets a briefing they can trust.
AgentDesk is built for teams that want agents to keep moving without losing the product narrative.
Live Observability
Track active runs, structured product observations, evidence, risks, blockers, and SLA breaches.
Living Dossiers
Auto-maintained product records that summarize current state, shipped outcomes, decisions, and open risks.
Stakeholder Briefings
Audience-specific briefing docs and decks that refresh automatically from product evidence.
MCP Server
Observability-first tools for runs, observations, evidence, dossiers, and briefing refresh.
REST + Export APIs
Programmatic access to runs, timelines, dossiers, briefings, and stakeholder-ready exports.
Enterprise Governance
Control export policy, retention, redaction, and stakeholder-facing role mapping in one place.
Connect in seconds
$ claude mcp add agentdesk # Your agents can now start runs, publish evidence, and keep dossiers current