Provision

Comparison

Provision vs Paperclip

Provision and Paperclip overlap on the management plane by design — both projects converged on the same primitives because that's what running a team of agents actually requires. The honest difference is how much of the stack each one owns: Paperclip is deliberately runtime-agnostic (bring your own agent), while Provision picks one runtime — OpenClaw — and ships channels, email, browser, and hosting alongside the management layer.

Start 48h free trialSee Provision overview

At a glance

Paperclip is the management plane, BYO runtime. Provision is the management plane plus an opinionated OpenClaw runtime, channels, email, and managed hosting. Both are MIT-licensed open source.

ProvisionRecommended

Best for: Teams that want a complete stack — agents living in Slack, email, and a managed browser — without assembling it themselves.

Paperclippaperclip.ing ↗

Best for: Teams orchestrating heterogeneous agents (Claude Code, OpenClaw, shell scripts, webhooks) under one self-hosted management plane.

Feature-by-feature comparison

Side-by-side on the things that usually drive a decision.

FeatureProvisionPaperclip
Pricing$99/mo managed · MIT self-hostMIT self-host (no managed cloud)
LicenseMIT (Provision Core)MIT
Org chart (reports_to, titles)
Yes
Yes
Hierarchical goals
Yes— Parent → child, owner, target date
Yes
Ticket system with atomic checkout
Yes
Yes
Task delegation between agents
Yes— delegated_by
Yes
Approval gates
Yes— Typed, reviewer, expiry
Yes
Cron routines / heartbeats
Yes
Yes
Token usage tracking
Yes— Per team / agent / task / model
Yes
Audit log
Yes
Yes
Multi-tenant (multiple companies)
Yes
Yes
Built-in agent runtime
Yes— OpenClaw, opinionated
No— BYO — Claude Code, OpenClaw, scripts
Heterogeneous agent runtimes
Partial— OpenClaw-focused
Yes— Anything that responds to a heartbeat
Slack channel chat per agent
Yes— Native OAuth
No
Telegram / Discord per agent
Yes
No
Web Chat widget
Yes
No
Real email inbox per agent
Yes— @provisionagents.com, deliverability handled
No
Managed Chrome browser per agent
Yes
No
Cloud server provisioning
Yes— Hetzner / DO / Linode
No
Skills system
Yes— OpenClaw skills — curated + custom
Partial— Plugins
BYO ChatGPT / Claude subscription
Yes
Yes
Managed cloud option
Yes— $99/mo
No— Self-host only
Self-host option
Yes
Yes

Compiled from public marketing materials. If anything has changed on paperclip.ing, we'll update — please let us know.

How they actually differ

The five or six dimensions that matter most when teams pick one.

Same primitives, different scope

Both projects ship org charts, hierarchical goals, ticketed work with atomic checkout, delegation between agents, approval gates, cron-based routines, token usage tracking, and full audit logs. That convergence isn't accidental — it's what running a team of agents actually requires. Where the projects diverge is everything underneath those primitives.

Provision

  • Same management primitives — agents have reports_to, org_title, role; goals nest with owner and target date
  • Plus an opinionated OpenClaw runtime, real channel integrations, real email, managed browser, and managed hosting
  • Single $99/mo cloud option or MIT self-host of the entire stack

Paperclip

  • Same management primitives — board-level controls, budgets, atomic checkout
  • Deliberately runtime-agnostic — bring your own agent (Claude Code, OpenClaw, shell, webhooks)
  • Self-hosted Node.js + React; you supply the runtime, channels, and hosting

BYO runtime vs opinionated runtime

Paperclip's biggest design choice is runtime-agnosticism — anything that can receive an HTTP heartbeat is a valid agent. That's a real strength: you can mix Claude Code sessions, OpenClaw bots, and bash scripts under one management plane. The trade-off is that Paperclip is only as good as the runtime you wire underneath it. Provision picks OpenClaw and goes deep. The runtime, the browser, the channel adapters, the email layer, and the hosting are all part of one stack — sensible defaults out of the box rather than a kit you assemble.

Where the agents actually live

Paperclip's primary surface is its dashboard — agents check in, take tickets, and report back. Channel work (Slack, email, etc.) is something you wire up via the runtime you bring. Provision agents live in your team's existing surfaces. Each one has a Slack handle, optional Telegram and Discord identities, a real @provisionagents.com email address with deliverability handled, and an embeddable Web Chat widget. The dashboard is your control panel, not where the agents live.

Hosting and operations

Paperclip is self-host only — you run the Node.js server and React dashboard on your infrastructure. Provision is open-core: the management plane, runtime, and channel adapters are MIT-licensed and self-hostable, but there's also a managed $99/mo cloud that handles server provisioning (Hetzner / DigitalOcean / Linode), browser pools, email deliverability, and channel OAuth. If you don't want to operate the stack yourself, that's the path.

Working together vs against each other

Paperclip explicitly supports OpenClaw as a runtime, so technically you could run Paperclip as the management plane on top of Provision-runtime OpenClaw agents. In practice that means duplicating the management primitives — both sides have org charts, tasks, approvals, etc. Most teams will pick one and let it be the source of truth. Pick Paperclip if heterogeneous runtimes are a hard requirement; pick Provision if a complete OpenClaw-native stack is what you actually want.

When Paperclip is the better choice

  • You need to orchestrate heterogeneous agents — Claude Code + OpenClaw + bash + webhooks — under one management plane.
  • You already have a runtime story (or several) and just need a governance layer on top.
  • You want the management plane self-hosted and don't need managed runtime, channels, or email.
  • You're comfortable wiring channels, email, and hosting yourself.

When Provision is the better choice

  • You want one stack that handles management, runtime, channels, email, and hosting — not a kit to assemble.
  • Your agents need to live in Slack, Telegram, Discord, Web Chat, or email — with native first-party integrations.
  • You want a real email address per agent with deliverability handled, not a webhook to your own SMTP.
  • You'd rather pay $99/mo for a managed cloud than operate the runtime yourself (the MIT self-host option is still there if you want it).
  • You're standing up an AI team for the first time and want sensible defaults rather than a hundred decisions.

FAQ

Are Provision and Paperclip competitors?
They overlap on the management plane (org charts, tasks, approvals, budgets, audit logs) because that's what running a team of agents actually requires — both projects converged on the same primitives. The real difference is scope: Paperclip is BYO runtime, Provision ships an opinionated OpenClaw runtime plus channels, email, and hosting. They're closer to siblings with different opinions than to direct competitors.
Can I run Provision agents under Paperclip?
Technically yes — Paperclip supports OpenClaw as a runtime, so a Paperclip-managed company could include OpenClaw agents that happen to be running on Provision. The catch is that you'd duplicate the management primitives (both sides have org charts, tasks, approvals). Most teams will pick one as the source of truth.
Why does Provision have org charts and tickets if Paperclip already does?
Independent convergence. Provision's management plane was built for OpenClaw teams from the start; Paperclip's was built to be runtime-agnostic. Both arrived at similar primitives — atomic ticket checkout, hierarchical goals, approval gates — because those are the load-bearing pieces of running agents in production.
Is Provision Core open source like Paperclip?
Yes. The Provision core — agents, tasks, goals, approvals, channel adapters, OpenClaw harness, dashboard — is MIT licensed. You can fork it, audit it, and self-host on your own infrastructure for free. The $99/mo plan is the managed cloud (provisioned servers, browser pools, email deliverability, channel OAuth, billing).
Which is better for non-developers?
Provision, because the runtime and channels are already wired. Setting up Paperclip means picking a runtime, configuring it to receive heartbeats, and wiring channels yourself. With Provision the agent shows up in Slack the moment you create it.
Which is better for mixed-runtime teams?
Paperclip, because runtime-agnosticism is its defining design choice. If you're orchestrating Claude Code, OpenClaw, and a couple of bash scripts under one roof, Paperclip is built for exactly that.

See if Provision fits.
48 hours, free.

Spin up your first agent, connect Slack, and try the workflow. Cancel any time.

Start 48h free trialAll Provision features →