Paperclip
paperclip.ing- Category
- AI Agents
- Rank
- No. 92Tools index
- Pricing
- Open Source
- Platform
- web · mobile · cli
- Type
- APP
- Builder
- @paperclipai
- GitHub
- 80.3k stars
- Latest release
- v2026.831.1
- Date
About
An open-source orchestration platform for managing teams of AI agents to run autonomous businesses. Create org charts, set budgets, define goals, and let AI employees handle operations from development to marketing.
What it does
Paperclip presents agent operations as persistent, ticket-based work. Scheduled heartbeats wake external agents through adapters, while a central dashboard records context, costs, approvals, reporting lines, and progress. Humans can inspect activity, change priorities, pause work, and intervene without owning every execution session.
Why it's ranked here
Paperclip addresses coordination problems that appear once several agents work concurrently. Atomic task checkout, persistent context, budget enforcement, approval gates, audit trails, and company isolation form a credible control plane rather than a thin launcher. The breadth is valuable, though it brings substantially more machinery than a simple agent runner.
What's good
Its strongest ideas concern operational control. Tasks retain conversations and goal ancestry across heartbeats. Budgets can stop agents at their limits. Humans can approve hires, pause agents, override decisions, and roll back revisioned configuration. Generic process and HTTP adapters sit beside integrations for several coding agents, so teams are not tied to one runtime.
Tradeoffs
Paperclip requires a server, database, agent adapters, and an organizational model, which may be excessive for one agent or occasional scripts. Agents still run externally, and richer status, cost, task, and log reporting depends on deeper integration. Several specification sections are explicitly drafts, while the task model says some documented behavior remains aspirational.
How to use it well
Use Paperclip when multiple long-running agents share goals, delegate work, consume measurable budgets, and need human oversight. Start with ticketed tasks and a small reporting structure, then add schedules, approval gates, and richer instrumentation. It does not supply the agents' reasoning or execution runtime, so pair it with callable command-line tools, processes, webhooks, or adapter plugins.
Technical notes+
The root package.json defines a private ESM pnpm workspace, requires Node.js 20 or newer, and supplies build, typecheck, Vitest, Playwright, database, release, smoke-test, and CLI scripts. packages/mcp-server/src/index.ts constructs a Paperclip MCP server over stdio and delegates operations to an API client. packages/db/src/index.ts exposes embedded PostgreSQL setup, migrations, backup, and restore facilities. packages/adapter-utils/src/index.ts exports runtime, billing, session, sandbox, redaction, and progress types and helpers. doc/CLI.md documents managed installation, onboarding, service control, diagnostics, profiles, JSON output, and control-plane API commands. doc/GOAL.md states that execution remains external and connects through adapters.
Observed
- License
- MIT License
- Primary language
- TypeScript
- Runtime and package manager
- Node.js 20 or newer with pnpm 9.15.4
- Application surface
- Node.js server with a React dashboard
- Interfaces
- CLI, REST API client operations, dashboard, and MCP server over stdio
- Database
- PostgreSQL support includes embedded setup, migrations, backup, and restore
- Service platforms
- macOS uses LaunchAgent; Linux and WSL2 use a systemd user unit when available
- Test surface
- Vitest, Playwright end-to-end tests, visual tests, smoke tests, and prompt evaluations are scripted
Read from README.md, package.json, packages/db/src/index.ts, packages/shared/src/index.ts, packages/mcp-server/src/index.ts, packages/adapter-utils/src/index.ts, packages/teams-catalog/src/index.ts, packages/skills-catalog/src/index.ts, packages/kv-demo-mcp-server/src/main.ts, packages/kv-demo-mcp-server/src/index.ts, packages/google-sheets-mcp-server/src/index.ts, doc/CLI.md, doc/GOAL.md, doc/SPEC.md, doc/TASKS.md.
What it can do
Create organizational charts for AI agent teams
Team structure requirements and roles → Organizational chart with defined AI agent hierarchies
Set and manage budgets for AI operations
Budget parameters and financial constraints → Budget allocation and spending controls for AI agents
Define and assign goals to AI agents
Business objectives and performance targets → Goal assignments and success metrics for AI agents
Orchestrate autonomous business operations
Business processes and operational requirements → Automated business workflows executed by AI agents
Deploy AI agents for software development tasks
Development requirements and specifications → Code, applications, or development deliverables
Execute marketing operations through AI agents
Marketing strategy and campaign parameters → Marketing campaigns and promotional content
Configure database and authentication systems
Database preferences and authentication requirements → Configured self-hosted platform infrastructure
Tags
Tech Stack
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.