8227485372
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work
> - Its trusted PR workflow builds the Runner's Rust binary in five
jobs: Typecheck + Release Registry, Build, Canary Dry Run, and the
`rust` and `vitest 2/2` lanes of Verify Paperclip Runner. The `Build`
job is the slowest check in almost every green run (461s, 489s and 515s
in runs 38180574768, 38182717052 and 38183149920 on 2026-10-11).
> - Each of those jobs restores master's `release-runner-v2` dependency
cache. That cache holds only third-party crates. `rust-cache` removes
workspace crates when it saves, and a fresh checkout invalidates cargo's
mtime fingerprints for workspace crates in any case. So cargo compiles
`paperclip-runner-core` in release mode from scratch on every run. The
job logs show one `Compiling` line and 28s to 49s for it.
> - 18 of the 239 commits on `master` in the seven days before this
change touched the crate's inputs. The other 221 compiled the same
sources to the same binary again.
> - This pull request adds a second cache that holds only that crate's
release artifacts (about 70 MB). The key is made from the compiler, the
checkout root and a hash of every input the build reads. The master Rust
lane writes it. A pull request whose merge tree has byte-identical
inputs restores it and marks the artifacts newer than the checkout, so
cargo keeps them. Every other pull request misses and compiles as
before.
> - The benefit is about 28s to 49s less on `Build`, which is the run's
critical path, and the same on Typecheck, Canary Dry Run and two Runner
lanes, for about nine of ten pull requests.
## Linked Issues or Issue Description
No GitHub issue exists for this. The description below follows the
enhancement template.
**What existing behavior does this improve?**
The time the trusted PR workflow spends to compile the Runner's own
crate on every run.
**Subsystem affected**
CI / build
**Current behavior**
Every PR job that builds the Runner binary compiles
`paperclip-runner-core` in release mode from scratch. The compile takes
28s to 49s on the `runs-on` fleet (one `Compiling paperclip-runner-core`
line, then `Finished release profile` in runs 38169436532 to 38183149920
on 2026-10-11). The shared Rust cache cannot help: it holds only
third-party crates, and a fresh checkout gives every source a new mtime,
which makes cargo discard the crate's fingerprints.
**Proposed behavior**
The master Rust lane in `release-verify.yml` saves the crate's release
artifacts under a key made from the compiler, the checkout root and a
hash of every input the release build reads. The PR jobs that build the
binary restore that entry read-only. On a hit, a script sets the mtime
of every restored artifact to now, so cargo sees every source as older
than the build and keeps the artifacts. `pnpm build` then prints
`Finished release profile` in under a second for that crate. On a miss,
cargo compiles the crate exactly as before.
**Reason and benefit**
About 28s to 49s off `Build`, the slowest check in almost every green
run, and the same off the Typecheck, Canary Dry Run and two Verify
Paperclip Runner lanes. The saving applies to every pull request whose
merge tree does not change the crate's inputs, which was 221 of 239
`master` commits in the week before this change.
**Breaking changes**
None. The checks are unchanged. A miss or an eviction costs the compile
time it costs today and nothing else.
**Additional context**
This is a workflow change. `pr.yml` calls `pr-trusted.yml@master`, so
this pull request's own run uses the current reader steps, and the
writer runs only on a master push. The saving becomes visible on the
first PR runs after the merge and after the next master push has written
the entry.
## What Changed
- `scripts/runner-release-crate-cache.mjs` (new): the one place that
defines the cached path list (`CACHED_PATHS`), the hashed inputs
(`HASHED_INPUTS`) and the files a restore must contain (`REQUIRED`). The
`scope` subcommand prints the platform, the `rustc --version --verbose`
release, commit and host, and a hash of the checkout root. The
`mark-current` subcommand sets the mtime of every restored artifact to
now and prints a `paperclip.runner-release-crate.v1` JSON line with the
file count and bytes. If the restored set lacks any of the seven files
cargo's freshness check reads or checks for (the fingerprint hash and
dep-info of each unit, the lib's rlib and rmeta, the binary under
`deps`), it removes the restored files, prints `action: discarded`
naming the missing file, and exits 0, so cargo compiles as on a miss.
- `.github/workflows/pr-trusted.yml`: after the `release-runner-v2`
restore in `typecheck_release_registry`, `build` and `canary_dry_run`,
and in the `rust` and `vitest 2/2` lanes of `verify_paperclip_runner`
(new matrix key `release_crate: true`), three steps: compute the key
scope, `actions/cache/restore` the crate under
`runner-release-crate-v1-<scope>-<hashFiles(...)>` with no
`restore-keys`, and run `mark-current` when `cache-hit == 'true'`. The
`hashFiles` list covers the toolchain pin, the manifests and lockfile,
the crate sources, the protocol tree and the three fixture trees the
test modules include (two under the Runner package and
`tests/runner-e2e/fixtures` at the repository root).
- `.github/workflows/release-verify.yml`: in the Rust lane of an
own-repository master push, after the debug warm-up: compute the key
scope, look the key up (`lookup-only: true`), and on a miss run `cargo
clean --release -p paperclip-runner-core` and `pnpm build:binary` (the
exact command the readers run, so the entry carries no test-profile
variants), then `actions/cache/save` the same path list under the same
key.
- `scripts/__tests__/runner-release-crate-cache.test.mjs` (new, 10
tests, in `test:release-registry`): the path list and hashed inputs are
the expected ones and exist; every `include_str!` and `include_bytes!`
in the crate sources resolves under a hashed input, whether its argument
is a literal or a `concat!` of literals, `env!("CARGO_MANIFEST_DIR")`
and macro metavariables, and any other form fails the test by name (this
guard found the `test/fixtures` tree during development and the
`tests/runner-e2e/fixtures` tree in review); the parser itself is tested
on a synthetic crate for each supported form and six rejected ones;
every PR reader restores after the dependency cache and marks before its
build command with the script's exact path list and key and no
`restore-keys`; the Verify lanes are gated by `release_crate`; the
writer runs after the checks with the lookup, the clean and build, and
the save under the readers' exact path list and key; the scope changes
with the compiler and the checkout root and survives a missing `rustc`;
`mark-current` makes every artifact newer than the checkout, leaves a
dependency's artifacts alone, discards a restore missing any one of the
seven required files (one case each, checking the named reason), and
still marks a restore missing any one of the seven files cargo does not
need.
## Review fixes
Greptile's two findings on the first commit were both valid and are
fixed in the second commit:
- **Incomplete caches looked usable.** `REQUIRED` checked four files,
but cargo's freshness check needs seven. Measured with cargo 1.97.1 on
the real crate (restore, touch the artifacts, delete one file, build):
without the lib's fingerprint hash, dep-info, rlib or rmeta it
recompiles the crate (24s to 27s); without the bin's fingerprint hash,
dep-info or the binary under `deps` it relinks (1.5s); without either
`.json` fingerprint record, either rustc `.d` file, the uplifted binary
or its `.d` it reports the crate fresh in 0.15s and recreates the
uplifted binary from `deps`, and the staging step then copies it.
`REQUIRED` now lists exactly the seven files, the glob for the binary
under `deps` excludes the `.d` beside it, and the test removes each of
the seven in turn (discarded, naming it) and each of the other seven in
turn (still marked).
- **Composed include paths escaped the guard.** The regex matched only
`include_str!("literal")` and skipped `include_str!(concat!(...))`,
which the crate uses four times (30 of 34 includes were checked). The
guard now parses every include argument and resolves the `concat!`
forms; one of them, in `provider_events.rs`, reads
`tests/runner-e2e/fixtures/native-completion/terminal-tool-carrier.json`
at the repository root, outside every hashed input. It is test-only, so
today's release binary does not read it, but the key is meant to cover
every include whatever its build, so `tests/runner-e2e/fixtures/**` (9
files, 80 KB) joins `HASHED_INPUTS` and the six key embeddings in the
two workflows. With the first commit's input list, the new guard fails
on exactly that include.
- `doc/RELEASE-AUTOMATION-SETUP.md`: a "Runner release crate cache"
section under the Runner verification dependency cache, with the key,
the writer, the readers and the trust boundary.
- `package.json`: the new test file in `test:release-registry`.
## Verification
- `node --test scripts/__tests__/runner-release-crate-cache.test.mjs`:
10 tests pass, three consecutive runs.
- The second commit, with the real toolchain on the real crate (`rustc
1.97.1`, cold build 2m20s, the crate's artifacts 74 MB): every source
touched to a fresh-checkout mtime and the archived artifacts restored,
cargo compiles (`Compiling paperclip-runner-core`, 24.8s and 25.1s, the
two controls); restored and marked by `mark-current`, `Finished release
profile in 0.16s` with no `Compiling`, and the same on a second build.
Then, for each of the fourteen files cargo wrote for the two units, one
at a time: restored, marked and that file deleted, cargo recompiles
without any of the seven `REQUIRED` files and reports fresh without any
of the other seven; restored with that file deleted and run through
`mark-current`, the seven required ones print `action: discarded` naming
the file and the others print `marked`. A restore holding only the seven
required files builds fresh in 0.15s, cargo recreates the uplifted
binary, and `stage-runner-binary.mjs` copies it into `dist/bin`.
- `node --test '.github/scripts/tests/*.test.mjs'`: 378 of 381 pass. The
3 failures are in `docker-canary-promotion.test.mjs` and fail the same
way on an unmodified `master` checkout in this sandbox (the stand-in
docker writes no `calls.jsonl`).
- `node --test scripts/__tests__/release-verify-workflow.test.mjs
scripts/__tests__/e2e-shard.test.mjs
scripts/__tests__/run-vitest-stable-shard.test.mjs`: 99 of 101 pass; the
2 failures spawn `pnpm`, which is not on this sandbox's PATH, and fail
the same way on `master`.
- Both workflows parse as YAML. No job has a duplicate step name or step
id.
- The mechanism, with the real toolchain (`rustc 1.97.1`, the package's
pin) on a copy of the crate: a cold `cargo build --release -p
paperclip-runner-core --bin paperclip-runnerd` (2m23s here). Then the
artifact set was archived with `tar`, removed, every source re-copied so
each had a new mtime, and the set restored:
- restored, not marked: `Compiling paperclip-runner-core`, 24.6s (the
control; this is what CI does today);
- restored and marked by `mark-current`: `Finished release profile in
0.15s`, no `Compiling`; a second build is the same;
- one Rust source edited after a marked restore: `Compiling`, 24.0s;
- one protocol schema edited after a marked restore: `Compiling`, 23.9s
(the protocol files enter cargo's dep-info as absolute paths and are
hashed into the key);
- the rlib removed from a restored set: `mark-current` prints `action:
discarded, reason: incomplete: the lib rlib is missing` and removes the
rest, and cargo compiles.
- Not run here: the workflow steps themselves. `pr.yml` pins
`pr-trusted.yml@master`, so this pull request's own `Build` job runs the
current reader steps, and the writer runs only on a master push. The
first PR runs after the merge and after the next master push show the
`Restore the Runner release crate` step's result and the
`paperclip.runner-release-crate.v1` line in their `Build` logs.
## Risks
- Correctness: a reader can only restore artifacts that master built
from inputs byte-identical to its own merge tree, because every input
the build reads is in the key, and the key has no prefix match. A pull
request cannot write. The trust boundary is unchanged: master writes, a
pull request restores master's entry.
- An input the build reads that is outside the hashed patterns would let
a stale artifact pass. The release dep-info lists the crate sources and
the protocol schemas and fixtures; all are hashed. The guard test
resolves every `include_str!` and `include_bytes!` in the crate, literal
or `concat!`, and fails if one lands outside the hashed inputs or uses a
form it cannot resolve, so a new include form cannot slip past it by not
matching a regex.
- GitHub-hosted runners check out under `/home/runner/work`, the
`runs-on` fleets under `/home/runner/_work`. Cargo records the included
protocol files as absolute paths, so an entry from one root is useless
under the other. The checkout root is in the key, so the other root
misses instead of restoring something cargo then discards.
- A cache that does not keep mtimes on restore changes nothing: the
artifacts are then already newer than the checkout, and `mark-current`
sets them to now in any case.
- Cost: one new entry of about 70 MB per distinct input set, written
only by the master Rust lane on a miss, plus one `cargo clean` and
`build:binary` (about 40s to 50s) on that lane when the inputs changed.
The readers pay one restore of about 70 MB (two to three seconds at the
fleet's observed restore speed) and a sub-second `mark-current`.
- An incomplete entry (a writer that saved a partial set) is discarded
by `mark-current` and costs one compile, the same as a miss.
## Model Used
- Claude (Anthropic), model id `claude-fable-5-1` (Claude Fable 5.1),
extended thinking, with tool use (shell, file editing, GitHub CLI, a
local Rust toolchain for the experiment) in a Paperclip agent session.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: devinfoley <139239+devinfoley@users.noreply.github.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Quickstart · Docs · GitHub · Discord · Twitter · Website
Sign up for the Paperclip Cloud waitlist →
Paperclip is the app people use to manage AI agents for work.
Open-source orchestration for teams of AI agents.
If OpenClaw is an employee, Paperclip is the company.
Paperclip is a Node.js server and React UI that orchestrates a team of AI agents to run a business. Bring your own agents, assign goals, and track work and costs from one dashboard. Choose models and harnesses per agent while keeping your team's tasks, skills, permissions, and history in one place.
It looks like a task manager. Under the hood: org charts, budgets, governance, goal alignment, and agent coordination.
Manage business goals, not pull requests.
| Step | Example | |
|---|---|---|
| 01 | Define the goal | "Build the #1 AI note-taking app to $1M MRR." |
| 02 | Hire the team | CEO, CTO, engineers, designers, marketers — any bot, any provider. |
| 03 | Approve and run | Review strategy. Set budgets. Hit go. Monitor from the dashboard. |
| Works with |
 OpenClaw |
 Claude Code |
 Codex |
 Cursor + Cloud |
 Gemini CLI |
 OpenCode |
 Pi |
 Hermes + Gateway |
 Grok Build |
 Kimi Code |
If it can receive a heartbeat, it's hired.
Custom processes, HTTP endpoints, and external adapter packages extend the roster. See the adapter overview for setup and capabilities.
Paperclip is right for you if
- ✅ You want to build autonomous AI organizations
- ✅ You coordinate many different agents (OpenClaw, Codex, Claude, Cursor) toward a common goal
- ✅ You have 20 simultaneous Claude Code terminals open and lose track of what everyone is doing
- ✅ You want agents running autonomously 24/7, but still want to audit work and chime in when needed
- ✅ You want to monitor costs and enforce budgets
- ✅ You want a process for managing agents that feels like using a task manager
- ✅ You want to manage your autonomous businesses from your phone
The four pillars
Four things have to work for an organization of AI agents to actually produce: the tasks, the org, the training, and the infrastructure. Paperclip is built around exactly those four pillars.
| Pillar | Built for | What it covers |
|---|---|---|
| Agentic Task Manager — Declare intent. Agents work. You verify the output. | Everyone, daily | Tasks, approvals & review gates · proactive agent coworkers · auditable routines & workflows · verify from diffs, screenshots & tests |
| Org Chart for Agents — Roles, permissions & boundaries for humans and agents. | Managers | Mixed human + agent org chart · responsibilities, delegation, specialization · governance: who can do what · scoped secrets & company boundaries · connection permissions & responsible-user identities |
| Agent Employee Training — Design, train & evaluate your AI employees. | Enablers | Skill Studio & shared org-wide skills · evals & saved test runs · active learning loops & quality metrics · performance reviews for agents · saved test inputs · skill version history & restore · reusable team templates |
| Agentic OS — The infrastructure that makes the work run. | IT & platform | Cross-provider runtime: any model, any agent · sandboxing, integrations & MCP servers · SSO, GRC, RBAC & cost controls · data privacy, internal trace collection, compounding data value · personal & shared app connections · run history & opt-in tracing |
Features
🔌 Bring Your Own AgentAny agent, any runtime, one org chart. If it can receive a heartbeat, it's hired. |
🎯 Goal AlignmentLink tasks and projects to your organization goals. Agents receive the goal context behind their work. |
💓 HeartbeatsAgents wake for assigned work, follow-up messages, or configured schedules. Delegation flows up and down the org chart. |
💰 Cost ControlCompany, agent, and project budgets. Track reported spend, get threshold alerts, and pause work at configured limits. |
🏢 Multi-OrganizationOne deployment, many organizations. Separate tasks, agents, permissions, and activity histories for each. |
🎫 Task ThreadsKeep conversations, plans, blockers, files, and run history attached to the work. Assign tasks to agents or people. |
🛡️ GovernanceConfigure review and approval stages, approve hires, and pause, reassign, or stop work when needed. |
📊 Org ChartHierarchies, roles, reporting lines. Your agents have a boss, a title, and a job description. |
📱 Mobile ReadyMonitor and manage your autonomous businesses from anywhere. |
🔗 Apps & ConnectionsConnect services such as GitHub, Notion, and Railway, or your own MCP server. Set gateway actions to Allowed, Ask first, or Off. |
👥 Shared Agents, Personal AccountsChoose who can use a connection and which agents can access it. Managed GitHub operations can use the account of the person directing the work. |
🧠 Skills & Skill StudioInstall or write shared skills, test them with saved inputs, inspect results, and restore earlier versions. |
📅 Scheduled RoutinesRun recurring work on a schedule or trigger it through an API or webhook. Each run has a task, an owner, and a history. |
📎 Artifacts & FeedbackFind the files and documents agents produce. Preview supported formats and leave comments on specific passages in documents. |
📦 Ready-Made TeamsPreview and install teams with roles, skills, projects, and routines. Choose their runtimes and make the setup your own. |
Experimental Agent Chat and chat/email connectors add conversations with agents in Paperclip and through configured services such as Slack, Discord, Telegram, and AgentMail. Enable the relevant instance settings to try them.
Problems Paperclip solves
| Without Paperclip | With Paperclip |
|---|---|
| ❌ You have 20 Claude Code tabs open and can't track which one does what. On reboot you lose everything. | ✅ Tasks are ticket-based, conversations are threaded, sessions persist across reboots. |
| ❌ You manually gather context from several places to remind your bot what you're actually doing. | ✅ Context flows from the task up through the project and company goals — your agent always knows what to do and why. |
| ❌ Folders of agent configs are disorganized and you're re-inventing task management, communication, and coordination between agents. | ✅ Paperclip gives you org charts, ticketing, delegation, and governance out of the box — so you run a company, not a pile of scripts. |
| ❌ Runaway loops waste hundreds of dollars of tokens and max your quota before you even know what happened. | ✅ Spend tracking, budget alerts, and automatic pauses help you control the cost of ongoing work. |
| ❌ You have recurring jobs (customer support, social, reports) and have to remember to manually kick them off. | ✅ Routines create assigned tasks on a schedule, with outputs and run history you can inspect. |
| ❌ You have an idea, you have to find your repo, fire up Claude Code, keep a tab open, and babysit it. | ✅ Add a task in Paperclip. Your coding agent works on it until it's done. Management reviews their work. |
Why Paperclip is special
Paperclip handles the hard orchestration details correctly.
| Atomic task checkout. | A single assignee and execution locks prevent competing runs from claiming the same task. |
| Persistent work context. | Tasks, comments, and documents stay in Paperclip. Supporting adapters resume saved sessions across runs. |
| Runtime skill injection. | Agents can learn Paperclip workflows and project context at runtime, without retraining. |
| Governance with rollback. | Approval gates are enforced, config changes are revisioned, and bad changes can be rolled back safely. |
| Accountable connections. | Human access, agent eligibility, and gateway action permissions are separate controls. Approve a call once or save a revocable rule. |
| Goal-aware execution. | Linked tasks and projects carry goal ancestry so agents see the "why," not just a title. |
| Portable company templates. | Export/import orgs, agents, and skills with secret scrubbing and collision handling. |
| Organization boundaries. | Company-scoped access checks keep each organization's work, agents, and activity separate within one deployment. |
What's Under the Hood
Paperclip is a full control plane, not a wrapper. Before you build any of this yourself, know that it already exists:
┌──────────────────────────────────────────────────────────────┐
│ PAPERCLIP SERVER │
│ │
│ ┌───────────┐ ┌───────────┐ ┌───────────┐ ┌───────────┐ │
│ │Identity & │ │ Work & │ │ Heartbeat │ │Governance │ │
│ │ Access │ │ Tasks │ │ Execution │ │& Approvals│ │
│ └───────────┘ └───────────┘ └───────────┘ └───────────┘ │
│ │
│ ┌───────────┐ ┌───────────┐ ┌───────────┐ ┌───────────┐ │
│ │ Org Chart │ │Workspaces │ │ Plugins │ │ Budget │ │
│ │ & Agents │ │ & Runtime │ │ │ │ & Costs │ │
│ └───────────┘ └───────────┘ └───────────┘ └───────────┘ │
│ │
│ ┌───────────┐ ┌───────────┐ ┌───────────┐ ┌───────────┐ │
│ │ Routines │ │ Secrets & │ │ Activity │ │ Company │ │
│ │& Schedules│ │ Storage │ │ & Events │ │Portability│ │
│ └───────────┘ └───────────┘ └───────────┘ └───────────┘ │
└──────────────────────────────────────────────────────────────┘
▲ ▲ ▲ ▲
┌─────┴─────┐ ┌─────┴─────┐ ┌─────┴─────┐ ┌─────┴─────┐
│ Claude │ │ Codex │ │ CLI │ │ HTTP/web │
│ Code │ │ │ │ agents │ │ bots │
└───────────┘ └───────────┘ └───────────┘ └───────────┘
The Systems
|
Identity & Access — Two deployment modes (trusted local or authenticated), human roles and permissions, agent API keys, short-lived run JWTs, company memberships, and invite flows. Responsible-user attribution follows work through delegation and supported managed connections. |
Org Chart & Agents — Agents have roles, titles, reporting lines, permissions, and budgets. Adapter examples match the diagram: Claude Code, Codex, CLI agents such as Cursor/Gemini/bash, HTTP/webhook bots such as OpenClaw, and external adapter plugins. If it can receive a heartbeat, it's hired. |
|
Work & Task System — Issues carry company/project/goal/parent links, atomic checkout with execution locks, first-class blocker dependencies, comments, documents, attachments, work products, labels, and inbox state. Search across work, review document revisions, and leave anchored feedback. |
Heartbeat Execution — DB-backed wakeup queue with coalescing, budget checks, workspace resolution, secret injection, skill loading, and adapter invocation. Runs produce logs, usage records, and adapter-specific session state. Bounded recovery handles supported failures and surfaces cases that need human action. |
|
Workspaces & Runtime — Project repositories and workspaces, optional isolated execution workspaces (git worktrees, operator branches), and runtime services (dev servers, preview URLs). Sandbox providers extend execution beyond the local host; availability depends on the configured environment and adapter. |
Governance & Approvals — Board approval workflows, execution policies with review/approval stages, decision tracking, budget hard-stops, and agent pause/resume/terminate. Configured task reviews govern completion; connection action approvals govern calls through the tool gateway. |
|
Budget & Cost Control — Token and cost tracking by company, agent, project, goal, issue, provider, and model. Scoped budget policies with warning thresholds and hard stops. Enforcement uses recorded spend; usage reporting and in-flight work can delay a stop. |
Routines & Schedules — Recurring tasks with cron, webhook, and API triggers. Concurrency and catch-up policies. Each routine execution creates a tracked issue and wakes the assigned agent — no manual kick-offs needed. |
|
Plugins — Instance-wide plugin system with out-of-process workers, capability-gated host services, job scheduling, tool exposure, and UI contributions. Extend Paperclip without forking it. |
Secrets & Storage — Company secrets and per-person secret values, encrypted credential storage, local or S3-compatible file storage, attachments, and work products. Secret references supply credentials to authorized runs without copying values into ordinary agent configuration. |
|
Activity & Events — Mutating actions, heartbeat state changes, cost events, approvals, comments, and work products are recorded as durable activity so operators can audit what happened and why. |
Company Portability — Preview, export, and import organization packages with agents, skills, and optional projects, routines, tasks, and attachments. Referenced secret values are omitted; review packages before sharing because plain environment values and local paths can remain. Packages share an operating setup; full-instance recovery uses backups. |
What Paperclip is not
| Not just a chatbot. | Conversations stay attached to tasks, plans, decisions, and outputs. Experimental Agent Chat can hand work off to assigned tasks. |
| Not an agent framework. | We don't tell you how to build agents. We tell you how to run a company made of them. |
| Not just a workflow builder. | Routines and experimental pipelines operate within an organization, with roles, goals, budgets, and governance. |
| Not a prompt manager. | Agents bring their own prompts, models, and runtimes. Paperclip manages the organization they work in. |
| Not limited to one agent. | Start with one agent and grow into a team with shared skills, delegation, and review. |
| Not only for code review. | Coding and PR review fit alongside research, operations, content, and other work. |
Quickstart
Open source. Self-hosted. No Paperclip account required. Follow the guided quickstart to set up your first agent.
Just ask your agent to install Paperclip
Share the installation guide with your agent.
Or install it yourself
With Node.js 24.11 or newer installed:
npx paperclipai@latest onboard --yes
The CLI runs from npm's cache; your instance configuration and data persist locally.
See the installation guide for managed installs, pinned versions, canary and git-ref installs, updates, rollback, service management, and uninstalling.
For an isolated manual test instance that is already initialized with a CEO
agent, use test-drive. It stays in the foreground, never installs a service
or creates a first task, and opens the browser only after setup succeeds:
ANTHROPIC_API_KEY=... npx paperclipai test-drive
OPENAI_API_KEY=... npx paperclipai test-drive --harness codex
OPENROUTER_API_KEY=... npx paperclipai test-drive \
--harness opencode \
--model openrouter/anthropic/claude-sonnet-4.5
Each run without --data-dir gets a unique, retained temporary directory; its
absolute path is printed at startup. Pass --data-dir to reuse one, or
--no-browser to leave the initialized instance unopened. When invoked from a
linked Git worktree, test-drive also enables task execution in that worktree.
See the test-drive guide for credential and
reuse behavior.
Troubleshooting: private npm registry
.npmrcIf this fails with an
E404forpaperclipai(or similar) and you use a private npm registry (for example GitHub Packages) via a global~/.npmrc,npxmay be resolvingpaperclipaiagainst that private registry instead of the public npm registry.Diagnostic:
npm config get registryWorkaround (cross-platform; force the public npm registry for this command):
npx --registry https://registry.npmjs.org paperclipai@latest onboard --yes
That quickstart path now defaults to trusted local loopback mode for the fastest first run. To start in authenticated/private mode instead, choose a bind preset explicitly:
npx paperclipai@latest onboard --yes --bind lan
# or:
npx paperclipai@latest onboard --yes --bind tailnet
If you already have Paperclip configured, rerunning onboard keeps the existing config in place. Use npx paperclipai configure to edit settings.
Or manually:
git clone https://github.com/paperclipai/paperclip.git
cd paperclip
pnpm install
pnpm dev
This starts the UI and API at http://localhost:3100. An embedded PostgreSQL database is created automatically — no setup required.
Requirements: Node.js 24.11+, pnpm 9.15+
Local Claude and Codex subscription sign-in also needs Python 3 and the corresponding provider CLI on the Paperclip host. The Docker image includes them.
Source development also builds the native Paperclip Runner when enabled (the self-hosted default). Install a Rust toolchain, or set PAPERCLIP_RUNNER_BINARY to a compatible prebuilt runner.
FAQ
Q: Is this project maintained or just slop?
A: Paperclip is maintained by the Paperclip team. We've merged over 2,700 pull requests.
Q: What does a typical setup look like?
A: Locally, a single Node.js process manages an embedded Postgres and local file storage. For production, point it at your own Postgres and deploy however you like. Configure projects, agents, and goals — the agents take care of the rest.
For remote access, use authenticated mode with a private-network bind such as Tailscale, or deploy the persistent server with Docker. See deployment modes and the Docker guide.
Q: Can I run multiple companies?
A: Yes. A single deployment can host multiple organizations with company-scoped data and access checks.
Q: How is Paperclip different from agents like OpenClaw or Claude Code?
A: Paperclip uses those agents. It orchestrates them into a company — with org charts, budgets, goals, governance, and accountability.
Q: Why should I use Paperclip instead of just pointing my OpenClaw to Asana or Trello?
A: Agent orchestration has subtleties in how you coordinate who has work checked out, how to maintain sessions, monitoring costs, establishing governance - Paperclip does this for you.
(Bring-your-own-ticket-system is on the Roadmap)
Q: Do agents run continuously?
A: Agents wake for assigned work and follow-up messages. Optional timer heartbeats let them check for work periodically; routines create recurring tasks on their own schedules. You can also connect externally running agents such as OpenClaw. A mention alone does not assign work or wake another agent.
Development
pnpm dev # Full dev (API + UI, watch mode)
pnpm dev:once # Full dev without file watching
pnpm dev:server # Server only
pnpm dev:mobile # Serve prebuilt UI on :3101 for phones/tablets (proxies /api → :3100)
pnpm dev:both # Run `pnpm dev` and `pnpm dev:mobile` together
pnpm build # Build all
pnpm typecheck # Type checking
pnpm test # Cheap default test run (Vitest only)
pnpm test:watch # Vitest watch mode
pnpm test:e2e # Playwright browser suite
pnpm db:generate # Generate DB migration
pnpm db:migrate # Apply migrations
pnpm test does not run Playwright. Browser suites stay separate and are typically run only when working on those flows or in CI.
See doc/DEVELOPING.md for the full development guide.
Roadmap
- ✅ Plugin system (e.g. add a knowledge base, custom tracing, queues, etc)
- ✅ Get OpenClaw / claw-style agent employees
- ✅ companies.sh - import and export entire organizations
- ✅ Easy AGENTS.md configurations
- ✅ Skills Manager, Skill Studio & Skills Store
- ✅ Scheduled Routines
- ✅ Better Budgeting
- ✅ Agent Reviews and Approvals
- ✅ Multiple Human Users
- ✅ Cloud / Sandbox agents (e2b, Cloudflare, Daytona, Modal, Novita, self-hosted Kubernetes)
- ✅ Artifacts & Work Products
- ✅ Deep Planning (planning mode, revisioned plans, plan approvals)
- ✅ Enforced Outcomes (watchdogs, recovery actions, review gates)
- ✅ MCP Tool Gateway & Apps (governed tool access)
- ✅ Secrets Manager with per-agent access
- ✅ Activity log & action attribution
- ✅ Self-healing runs & automatic recovery
- ✅ Agent evals & feedback
- ✅ Connected Apps
- ✅ Personal & Shared AI Accounts
- ✅ Shared Agents Use Personal GitHub Identities
- ✅ Skill Version History & Restore
- ✅ Document Comments & Revision History
- ✅ Company-Wide Search
- ✅ Multi-Model & Multi-Harness Teams
- 🟡 Memory / Knowledge
- ⚪ MAXIMIZER MODE
- ⚪ Work Queues
- ⚪ Self-Organization
- ⚪ Automatic Organizational Learning
- 🟡 Agent Chat
- 🟡 Cloud deployments
- ⚪ Desktop App
- ⚪ Bring-your-own-ticket-system (Asana / Linear / Jira as on-ramps)
This is the short roadmap preview. See the full roadmap in ROADMAP.md.
Community & Plugins
Find Plugins and more at awesome-paperclip
Observability
Paperclip ships with opt-in OpenTelemetry auto-instrumentation for the server (traces only). It activates when OTEL_EXPORTER_OTLP_ENDPOINT is set and supports grpc, http/protobuf, and http/json via the standard OTEL_EXPORTER_OTLP_PROTOCOL env var. @opentelemetry/api is a normal server dependency; the SDK, auto-instrumentation, and exporter packages are optional peer dependencies — install them only if you want tracing. See doc/observability.md for install commands and the full env-var reference.
Paperclip also ships with opt-in Sentry error monitoring for the server and the browser. Set SENTRY_DSN_FRONTEND to activate it for the browser and SENTRY_DSN_BACKEND to activate it for the server — each variable is optional, and the legacy SENTRY_DSN variable still works as a fallback for either component. The supported server SDK version is @sentry/node@10.71.0; it is an optional peer dependency for the server, so install it only if you want error monitoring. The browser SDK, @sentry/browser, is pinned to the same exact version. See doc/observability.md for the install command, the privacy settings, and the full default capture set.
Telemetry
Paperclip collects anonymous usage telemetry to help us understand how the product is used and improve it. No personal information, issue content, prompts, file paths, or secrets are ever collected. Private repository references are hashed with a per-install salt before being sent.
Contributors changing emitted telemetry events should follow the Telemetry Data Contract. For proposed first-party events that are not in the generated contract yet, follow Telemetry Workflow.
Telemetry is enabled by default and can be disabled with any of the following:
| Method | How |
|---|---|
| Environment variable | PAPERCLIP_TELEMETRY_DISABLED=1 |
| Standard convention | DO_NOT_TRACK=1 |
| CI environments | Automatically disabled when CI=true |
| Config file | Set telemetry.enabled: false in your Paperclip config |
Contributing
We welcome contributions. See the contributing guide for details.
Community
- Discord — Join the community
- Twitter / X — Follow updates and announcements
- GitHub Issues — bugs and feature requests
- GitHub Discussions — ideas and RFC
License
MIT © 2026 Paperclip Labs, Inc
Star History
Open source under MIT. Built for people who want to get work done, not babysit agents.