Goes live November 17, 2026

It appears on the home page and opens for votes when its launch week starts.

Arbituro

Arbituro

governed AI agents for real work

FreeAIWorkflow automationAI Agents5 impressionslaunches Nov 171 comment

Comments

>log in to comment
  • Steven Innes[maker]· 16h ago

    For teams that need AI to draft work but keep full control: Arbituro lets agents turn briefs into reports, memos and decks while every action requires human approval and is logged. It scopes access, shows an approval queue and records acceptance criteria. Would love your thoughts on the workflow!

Arbituro, by Writan Labs: work management with AI agents and governance built in. Scoped access, human approval for anything sent out, and a full audit trail. Agents turn briefs into reports, memos and decks, tidy your backlog and draft your plan. You have the final say. The brief, its acceptance criteria and the delivered deck, with the agent beside it. Agents in Arbituro reach only what a person grants. Nothing they change or send happens until a person approves it, and every step is on the record. Before work starts, an agent suggests acceptance criteria and you agree them. Each one says who or what decides it, and every decision is recorded and never overwritten. The delivered deck waiting for review, and the acceptance criteria it is judged against, each naming who checks it.

More AI tools →80 free dev tools →

  • Governed AI agents with human approval
  • Acceptance criteria recorded and never overwritten
  • Unified approval queue for all tasks
  • Agents like Scout and Sage handle backlog and work
  • Free to start for up to three people

Arbituro provides governed AI agents that generate reports, memos, decks and manage backlogs with human approval and full audit trails.

for
Teams that need AI-generated work but require strict governance and approval workflows.
pricing
freemium · free trial
license
MIT
writan-labs/arbituro-mcp 0 0updated 2 days ago
works withMicrosoft 365Google WorkspaceSlackGitHub

Key features

8 features of Arbituro
  • Governed execution — Agents can only act on resources granted by a person and all actions require human approval.
  • Acceptance criteria — Agents propose acceptance criteria before work starts and record every decision.
  • Approval queue — All pending approvals, reviews and questions appear in a single queue with wait times.
  • Audit trail — Every agent action is logged and exportable as CSV for full traceability.
  • Backlog management — Agents suggest duplicate removal, job splitting and reassignment with reasons for each change.
  • Planning assistance — Agents draft schedules for undated jobs based on priority, due dates and workload.
  • Coding agents — Integrates Claude Code, Cursor and VS Code to handle coding tasks with audited pull-requests.
  • Integrations & access control — Connects to Microsoft 365, Google Workspace, Slack and GitHub with granular read/write permissions.

Use cases

  • Generate a market report deck that matches brand style and includes source citations.
  • Automate backlog cleanup by having agents propose duplicate removal and job splitting.
  • Draft project schedules and have managers approve dates before tasks move forward.
  • Create code changes via a coding agent that submits pull requests meeting predefined criteria.

Arbituro vs alternatives

Arbituron8nLangGraphLlamaIndexMastra
Best forGoverned AI agents for enterprise workNo-code workflow automationStateful agent frameworksLLM data indexingTypeScript agent development
PricingFreemiumSubscriptionFreeFreeFree
DevHunt upvotes10000
LaunchedNov 2026————
  • Arbituro vs n8n: n8n offers generic workflow automation without built-in governance or AI-generated deliverables.
  • Arbituro vs LangGraph: LangGraph is a framework for building controllable AI agents, not a hosted managed service like Arbituro.
  • Arbituro vs LlamaIndex: LlamaIndex focuses on data indexing for LLM apps, whereas Arbituro provides end-to-end work management with approvals.
  • Arbituro vs Mastra: Mastra is a TypeScript framework for AI agents, requiring developer effort; Arbituro is a ready-to-use platform with UI and governance.

Arbituro FAQ

Can an agent send an email or post a message on its own?+

No. Every email, chat message or calendar invitation must be approved by a designated person before it is sent.

What can an agent access?+

Only the brief, sources, documents shared by a project admin and any integrations explicitly granted; web search is disabled by default.

Can we require approval before work moves on?+

Yes. Any board column can be set to require approval before a job advances, enforcing segregation of duties.

How is agent usage billed?+

Usage is metered in credits against the plan’s monthly allowance; you can cap spend per agent or per job.

Which AI models are supported?+

Models from OpenAI, Anthropic, Google, Mistral and others are available, selectable per agent without managing API keys.

Can I control who sees what?+

Project visibility is limited to members, and roles (owner, admin, contributor, viewer) define edit permissions.

Summarized by DevHunt from arbituro.com · Nov 15, 2026. Details may change; check the official site.