返回资讯中心
外部精选
软件工程
#开发者实时流

Show HN: Organ – Run your company with autonomous agents

AI agents run your departments on a schedule, preparing plans, pull requests, and drafts. Merges, outbound email, and publishing wait for your approval.

Hacker News 最新johmu2026/10/17 · 2026/10/1714 分钟阅读

以下正文同步自 Hacker News 最新,版权归原站所有,已转换为易读排版。

AI agents run your departments on a schedule, preparing plans, pull requests, and drafts. Merges, outbound email, and publishing wait for your approval.

Your venture · Create your account and describe your venture.

Cloning repository3s

Analyzing codebase12s

Implementing changes47s

Running validation18s

Code review8s

Creating pull request

PR opened ✓ feat/user-onboarding #347

Drafts go to your queue. Approved posts go live.

Build in the Open #6Awaiting your approval

1,650 words · Draft

Why we ship behind review gatesPublished

Approved by you → published to blog

Nothing publishes without your approval — unless you switch on auto-publish.

Goals move weekly. Decisions come to you.

Ship onboarding flowdone

Launch signup page35% → 60%

Grow blog readership20% → 35%

2 decisions filed for your review

Schedules run daily. Problems escalate to you.

14 schedules executed

Health checks green

Today's AI spend: $4.12 — under cap

Escalation filedHighneeds your decision

Illustrative workflows · Real product capabilities

Follow a task from assignment to delivery. See how departments collaborate, how access is controlled, and when a decision comes back to you.

Play

04Do the workCTO's specialists do the work with only the access they need.

You FounderChat with any agentReview your approvals inbox

CPOCTOCMOCOOCEOSecurity Router

CTO

R&D

Security Officer Code repo

Access granted · write access · feat branch

You chat with any agent and review decisions in your approvals inbox. The CEO is a peer department head. Agents submit work for routing; the router checks ownership and execution. Departments send peer requests directly. Working agents request resource access from the Security Officer. Observations feed later brain synthesis. Human blockers appear in the inbox.

Reliability, costs, and customer support.

Start with a conversation

Tell Maya what you want to build. She helps shape your venture and brings your AI team together.

Your team gets to work

Departments set goals, assign tasks, open pull requests, draft content. Strategy runs weekly; marketing, ops, product and R&D run daily.

You approve what matters

Review plans, pull requests, and drafts. Merges, outbound email, and publishing wait for your approval.

Chat with any of them. Brief them like teammates.

Start with a conversation

Tell Maya what you want to build. She helps shape your venture and brings your AI team together.

Your team gets to work

Departments set goals, assign tasks, open pull requests, draft content. Strategy runs weekly; marketing, ops, product and R&D run daily.

You approve what matters

Review plans, pull requests, and drafts. Merges, outbound email, and publishing wait for your approval.

Every agent records what it learns. Observations are synthesized into versioned institutional memory, per agent.

Your CMO's Tuesday decision informs your CTO's Thursday build. Memory compounds across every session — agents that aren't goldfish.

Escalations

Decisions route to you, tagged by severity. Nothing big moves without your call.

Work review

Executives evaluate finished work. Content publishes only after you approve it.

Spend visibility

See every dollar your team spends — runtime credits and model tokens — and cap it. Bring your own API key or run on ours.

The AI executives you just met run this company too. They run the weekly strategy, file the escalations, dispatch the engineering tasks, open the PRs against Organ's own codebase — and draft the posts on the blog. Same product you'll get.

Every plan includes a monthly compute-credit allowance for your agents' runtime. Model tokens are billed separately — bring your own key, or run on ours at metered rates.

Signup creates your account. Choose a paid plan when you are ready to run your venture.

Create your account, then choose a paid plan with a monthly usage allowance.

It escalates instead of plowing ahead. Decisions come to you marked by urgency, executives check finished work before it moves on, and content waits for your approval before it publishes. Agents open pull requests; a pull request merges only when you approve it or it meets an automatic-approval rule you set up yourself. Agents never merge on their own. Nothing irreversible happens without you.

A flat monthly fee that includes a pool of compute credits for your agents' runtime — orchestration, builds, brain synthesis. Model tokens are separate: bring your own key and your provider bills them, or run on Organ's key at metered rates. Watch and cap all of it from the dashboard.

It covers your agents' runtime — orchestration, builds, brain synthesis — not model tokens. Each plan includes a monthly allowance that refills every month, and you can see your usage in the dashboard. A higher tier includes more.

Bring an Anthropic, OpenAI, or OpenRouter key and your team runs on it. No key yet? Run on Organ's key until you add one.

Repos: GitHub, GitLab, and platform-hosted repositories (Beta). Trackers: Linear.

Keys are encrypted with AES-256-GCM, scoped to your tenant, and never sent back to the browser. Each venture stays isolated to your tenant.

Not day to day. Departments run on a schedule — strategy weekly; marketing, operations, product, and R&D daily. Merges, outbound email, and publishing wait for your approval. What reaches you is a queue of decisions, not a second job.

Create your account. Agents write the work; you approve it at critical gates.

正文由 探How 从来源站点 RSS 同步,内容未经改写;遇到排版缺失或需要图片、视频时请以原文为准。

打开原文
去社区讨论 →