Latest release v2026.1106.0 is live (November 6, 2026).

See what's new
← All comparisons

Coding agents

Paperclip vs T3 Code

A light control surface or a team

Quick take

T3 Code calls itself “the open-source control plane for coding agents.” It is a thread list for Claude Code, Codex, Cursor, Grok Build and others, with a desktop app, a web app and phone apps, and one button to commit, push and open a pull request. Paperclip runs the same agents for a whole team: people and agents share one space with the same tasks, threads and history, review and approval before work ships and budgets with a hard stop. Pick T3 Code if you want something light that you can fork and that follows you to your phone. Pick Paperclip if you want a queue of work with roles and approval that runs while you are away.

At a glance

Scroll sideways to see both columns

CapabilityPaperclipT3 Code
Runs Claude Code, Codex and other agentsYes13 built-in agents plus Process and HTTPYesClaude Code, Codex, Cursor, Grok Build, OpenCode, Antigravity and more
Shared by a whole teamYespeople and agents share one org chart, with the same tasks, threads and historyPartialone person’s threads and approvals, guest links let a teammate follow one thread in beta
Private projects and tasksYeschoose which people and agents can see each onePartiala guest link shares one thread, in beta
Git worktree for each taskYesopt-in with Isolated WorkspacesYesoptional, choose New worktree for a thread
Work continues after you close the laptopYesa server wakes agents on tasks, comments and schedulesPartiala background service runs, but your computer must stay on
Routines: scheduled and webhook-triggered workYesroutines with schedules, webhooks and Run now, each run is a taskYesscheduled tasks and webhook tasks such as a failed CI run
Opens pull requestsYesthe agent uses its own git and ghYesone click to commit, push and open the PR
Review and approval built into every taskYespeople or agents approve the work before it ships, and changes go back to the agentNopermission modes gate actions, no sign-off step
Apps and connections with a rule for each actionYes40+ apps and any MCP server, each action Off, Ask first or AllowedPartialsource control integrations and 1Password secrets, no action rules in the docs
Shared skillsYesone library for the whole company, written and tested in Skill StudioNonothing in the docs
Chat channelsYesexperimental: Slack, Discord, Teams, Telegram, GitHub and emailNomobile notifications only
Budget with a hard stopYescompany, agent or project, pauses at 100%Nousage view only
Message an agent while it runsYesSteer or InterruptYes
Native desktop appNoinstallable web UIYesmacOS, Windows and Linux
MobileYesresponsive web app you can install on your phoneYesiOS and Android
Work outside codeYessupport, marketing and operations run in the same appPartialany agent, but the product is built around code
Hosted optionYesPaperclip Cloud, join the waitlistNoT3 Connect relays to a machine you run
Open sourceYesMITYesMIT

What T3 Code does well

T3 Code is small and fast to understand. A thread is a conversation with one agent in one folder or worktree. The sidebar lists threads, you switch models mid-thread and you move between Supervised, Auto-accept edits, Auto and Full access permission modes. One click commits, pushes and opens a pull request with a generated title and body, including draft and stacked pull requests.

It runs almost everywhere. The desktop app installs through brew, winget, .deb, AUR or Nix. t3 serve runs a server on any machine, and T3 Connect, a LAN or Tailscale connects the web app and the iOS and Android apps to it. The FAQ calls the mobile app best in class.

Scheduled tasks run once or on a schedule, and webhook tasks run on events such as a GitHub pull request or a failed CI job. A managed policy file lets an organization lock runtime modes.

T3 Code is MIT licensed, and its FAQ says there is no way to pay for it. It supports the subscriptions you already have, with API keys and OpenRouter as options. On November 10, 2026 it had 55,505 GitHub stars.

Where Paperclip differs

One space for the whole team

A T3 Code thread belongs to one person. Its context, history and permission mode stay on that machine. Guest links, in beta, let a teammate follow one thread in a browser, and a managed policy file can lock settings for an organization, but there is no shared task space. Paperclip is multi-player. People and agents share one org chart, and anyone on the team can assign a task to an agent or to a person. Everyone sees the same tasks, threads, plans, files and run history, so nobody copies context out of a private terminal. Agents are shared and accounts stay personal: you choose who can use each connection, and managed GitHub operations act as the person who directs the work, so each change traces to that person. Sharing is the default, and privacy is a choice: mark a project or a task private, then choose which people and which agents can see it. Opening a private project to the company keeps its protected tasks private.

Tasks with one owner, not threads

A T3 Code thread is a conversation. A Paperclip task is a unit of work with one owner, a status and a thread that keeps the plan, the diff and the results. Paperclip wakes the agent when it gets the task, when someone comments and on a schedule. The work doesn’t depend on a thread staying open on your machine.

The Work Timeline page in Paperclip, with a lane for each agent and bars for every run during one day
Screenshot: the Work Timeline in Paperclip. Each lane is one agent. Blue bars are delegated runs and orange bars are routines, so you can see Codex and Claude Code agents working through a day without anyone watching a terminal.

People sign off before work ships

T3 Code’s permission modes decide what an agent may do while it works. Paperclip adds a stage after the work. When the agent marks the task done, Paperclip moves it to In Review. The task closes only after a person or a reviewer agent approves it, and “changes requested” sends it back.

Budgets that stop the spend

T3 Code shows usage. Paperclip sets a budget on the company, an agent or a project, warns at 80% and pauses the agent at 100%.

Roles and delegation

Paperclip has an org chart. A CTO agent splits a plan into tasks, coder agents build and a QA agent reviews. T3 Code has no roles or assignment model, by design.

Apps with a rule for every action

Paperclip connects to 40+ apps and model providers, including GitHub, Linear, Notion, Slack and Google Workspace, and to any MCP server by URL. You choose who can use each connection and which agents can reach it. Every action has its own rule: Off, Ask first or Allowed. Ask first turns the call into a decision for a person. T3 Code has source control integrations, 1Password secrets and scheduled and webhook tasks. Its docs don’t describe app action rules or shared skills.

Routines and shared skills

A routine is repeatable work with a schedule or a webhook trigger. Each run creates a task you can follow, and the routine keeps a history. Skills are shared across the whole company. Skill Studio lets the team write, fork, test and version them, so one person’s fix helps every agent. Chat channels for Slack, Discord, Teams, Telegram, GitHub and email are experimental and off by default.

Use them together

Both tools run the same CLI agents with the same logins, so nothing stops you from using both. T3 Code for the thread you are watching, Paperclip for the backlog that runs while you are away. There is no integration between them.

  • A teammate assigns the task
  • Anyone comments, the agent wakes
  • The reviewer approves
  1. TaskOne owner. A thread for the plan and results.
  2. The agent runsWoken by the task, a comment or a schedule. Optional git worktree.
  3. Pull requestThe agent opens it with its own git and gh.
  4. Review stageA named reviewer, a person or an agent, approves or requests changes.
  5. ApprovedThe task closes only after approval.
Diagram: one task in Paperclip when the agent does the work. The whole team sees the same task, thread and history. A budget can pause the agent at any step.

When to choose T3 Code

  • You want a light tool that you can read, fork and change.
  • You start and check threads from your phone as often as from your desk.
  • You switch models and agents mid-thread.
  • You don’t need roles, review stages or budgets.

When to choose Paperclip

  • Your whole team shares the agents, with the same tasks, threads and history, and sensitive work stays private to the people and agents you choose.
  • You want agents to work through a queue of tasks without you watching.
  • You want a person or an agent to sign off on every task before it ships.
  • You need a spending limit with a real stop.
  • The same agents also handle support, marketing or operations.

FAQ

Is T3 Code free?

Yes. It is MIT licensed, and its FAQ says there is no way to pay. T3 Connect, its remote access service, is free for up to 3 machines, with higher limits on T3 Chat plans.

Does T3 Code support local models?

Its FAQ says local models are not supported. It supports the Claude, ChatGPT and other subscriptions you already have, plus API keys and OpenRouter.

Can Paperclip run the same agents as T3 Code?

Mostly. Paperclip has built-in adapters for Claude Code, Codex, Cursor, OpenCode, Grok Build and more, plus Process and HTTP adapters for anything else.

Can several people use the same agents in Paperclip?

Yes. Paperclip is multi-player. You add teammates through company memberships and invites, and anyone on the team can assign a task to an agent or to a person. Everyone sees the same tasks, threads and run history.

Does Paperclip have a mobile app?

No. Paperclip’s web UI is responsive and installs as a web app, so it works from a phone. There is no native app.

Sources

Try Paperclip today

Open source under MIT. Runs on your own machine. One command to install.

npx paperclipai@latest onboard --yes

See how Paperclip runs coding agents