Trooper is an AI workforce platform powered by OpenClaw technology. It lets you deploy multiple AI employees (agents) that work autonomously — making GitHub commits, sending emails, browsing the web, and executing complex multi-step tasks.

What are OpenClaw Skills?

OpenClaw Skills are installable capabilities for AI agents. With 3,000+ skills available, they extend agent functionality across coding, browser automation, data analytics, DevOps, testing, and more. Install via: npx clawhub@latest install .

How does GitHub integration work?

Trooper AI agents connect directly to your GitHub repositories. They can make real commits, create pull requests, review code, manage branches, and resolve merge conflicts — all autonomously.

Can multiple AI agents work together?

Yes, Trooper supports multi-agent collaboration. Multiple AI agents can work on the same project, communicate with each other, delegate tasks, and orchestrate complex workflows together.

trooper

[01]Unit brief

Trooper for Coding
Use Codex & Claude Code together.

Delegate to Claude Code, Codex, Cursor, and your own agents from one Trooper harness. Your subscriptions, your repos, one unified task board with traced diffs and PR workflows.

Deploy unit Talk to founder

app.trooper.so

Channels

Direct messages

Do anything with AI…

Columns

List

🎯TODO

Nothing here yet

Drop a task here or add one with +

⏳IN PROGRESS

Nothing here yet

Drop a task here or add one with +

📝HUMAN REVIEW

Nothing here yet

Drop a task here or add one with +

🎉COMPLETED

Nothing here yet

Drop a task here or add one with +

[02] Overview

Multi-agent coding ops

Trooper does not resell model access. You bring Claude, OpenAI, or any provider subscription — we bring the harness: routing, permissions, logs, and shared memory across every agent on your roster.

Claude Code hardens a parser, Codex dedupes ETL rows, and your in-house agent drafts release notes — all in parallel, all on one board.

Every edit, test run, commit, and PR open is traced in the ticket thread. Senior engineers approve merges; agents handle the grind.

Multi-agent harness

Run Claude Code, Codex, and custom agents side-by-side — each on its own task with live diffs and activity logs.

Unified task board

In progress, backlog, and done — one board for every coding agent. Pick the right soldier for the job, not the right browser tab.

BYOA subscriptions

Your API keys and IDE subscriptions stay yours. Trooper routes work, enforces permissions, and keeps shared org memory.

[03] Playbooks

Build agents that run your playbooks.

Reusable coding runbooks — PR review, hotfix, and release train — traced on the board with GitHub, Slack, and harness agents.

GitHub webhook

Trigger

PR opened

Review harness

Inline comments

Notify #eng-prs

Merge gate

A/B test: comment style+18% approval rate

[04] Capabilities

How this unit runson Trooper.

Traced tickets, live artifacts, and harnesses that match the work — not generic placeholders.

[04]MULTI-AGENT HARNESS

Your subscriptions. Our harness.

Run Claude Code, Codex, and OpenCode side-by-side on one ticket — parser patch, ETL dedupe, and regression tests with live tool traces and diff stats.

▸Three harness lanes with provider logos and branch names
▸exec, apply_patch, and write_file traced per agent
▸BYOA — your keys, Trooper routes work and enforces permissions

trooper · ticket #1 · harness

ENG · parser hotfixIn progress

Fix invoice parser null rows

acme-billing3 harnesses · BYOA

Codex

Leo · fix/parser-null-rows

+5 −1

src/parser.ts

execnpm test — repro failure

apply_patchfilter empty CSV rows

execnpm run test:integration

OpenCode

Leo · fix/etl-dedupe

+3 −0

etl/dedupe.ts

apply_patchignore empty row ids

execdedupe.integration.test

Claude Code

Ren · fix/parser-null-rows

+12 −0

tests/parser.test.ts

write_fileempty CSV regression case

execvitest parser.test.ts

Subtasks 4/6·PR #418 queued

Live trace

[05]CANVAS + CI

Diffs, CI logs, and PR bundles on Canvas.

Open the ticket Canvas for parser.ts, etl/dedupe.ts, integration CI output, and PR #418 body — the same artifacts from the live demo, stacked for review.

▸Unified diffs with line numbers and +/− counts
▸Full integration log — 13 tests passed, CI green
▸PR body with summary, changes, and merge gate

Canvas4 artifacts · desktop workspace

src/parser.ts.diff

src/parser.ts+5−1

12− const rows = raw.split('\n').map(line => line.trim());

13+ const rows = raw.split('\n')

14+ .map(line => line.trim())

15+ .filter(Boolean);

16 return rows.map(parseRow).filter(Boolean);

etl/dedupe.ts.diff

etl/dedupe.ts+3−0

44+ if (!row.id?.trim()) return false;

45 const key = `${row.id}:${row.batchId}`;

46 if (seen.has(key)) return false;

logs/ci-integration.log

$ npm run test:integration

✓ parser.integration.test.ts (8 tests) 2.1s

✓ etl/dedupe.integration.test.ts (5 tests) 1.4s

Test Files 2 passed (2)

Tests 13 passed (13)

CI · green · ready for PR #418

pull-requests/418-body.md

PR #418 — fix(parser): skip empty invoice rows

Summary

Production CSVs include blank lines that caused the invoice parser to drop valid rows.

Changes

src/parser.ts — filter empty rows after trim

etl/dedupe.ts — ignore rows with empty ids

tests/parser.test.ts — empty-row regression

Awaiting merge approval from @Vaibhav

Canvas workspace4 artifacts · organized

[06]ORG CONTEXT

Branch rules and reviewer prefs persist across agents.

AGENTS.md, eslint rules, CODEOWNERS, and merge gates load on every coding mission — Codex, Claude Code, and OpenCode share the same org memory.

▸Default branch, test commands, and reviewer routing
▸Lint and import-order conventions enforced in src/
▸CI required + human approval before merge

trooper · org memory

Loaded on every coding missionPersisted

Branch rules, lint config, and reviewer prefs — no re-briefing each session.

AGENTS.md

default_branchmain — no direct pushes

reviewer@Vaibhav on parser/*

test_cmdnpm test && npm run test:integration

.eslintrc.cjs

no-consoleerror in src/

import/orderenforced · autofix on save

@typescript-eslintstrict mode

CODEOWNERS

src/parser/*@Leo @engineering

etl/*@Leo

*@engineering

merge-gates.yml

ci_required✓ integration suite green

approvals✓ 1 human before merge

atomic_checkout✓ one agent per branch

4 files · synced across Codex, Claude Code, OpenCodeMemory hit · 0ms re-brief

[07]DESKTOP CANVAS

Parser hotfix bundle on Canvas.

Four artifacts from one mission — parser diff, ETL patch, green CI, and PR body arranged for merge review.

▸Diffs, logs, and PR copy visible at once
▸Live review mode with cursors and comments
▸Drag title bars to organize before approval

Canvas4 artifacts · live review

Review mode

src/parser.ts.diff

src/parser.ts+5−1

12− const rows = raw.split('\n').map(line => line.trim());

13+ const rows = raw.split('\n')

14+ .map(line => line.trim())

15+ .filter(Boolean);

16 return rows.map(parseRow).filter(Boolean);

etl/dedupe.ts.diff

etl/dedupe.ts+3−0

44+ if (!row.id?.trim()) return false;

45 const key = `${row.id}:${row.batchId}`;

46 if (seen.has(key)) return false;

logs/ci-integration.log

$ npm run test:integration

✓ parser.integration.test.ts (8 tests) 2.1s

✓ etl/dedupe.integration.test.ts (5 tests) 1.4s

Test Files 2 passed (2)

Tests 13 passed (13)

CI · green · ready for PR #418

pull-requests/418-body.md

PR #418 — fix(parser): skip empty invoice rows

Summary

Production CSVs include blank lines that caused the invoice parser to drop valid rows.

Changes

src/parser.ts — filter empty rows after trim

etl/dedupe.ts — ignore rows with empty ids

tests/parser.test.ts — empty-row regression

Awaiting merge approval from @Vaibhav

Canvas workspace4 artifacts · organized

[10] Field ops

How Trooper runs your code unit

From invoice parsers to release notes — agents execute full workflows, not single prompts.

Parallel patch missions

Three agents, three files, three test suites — all tracked with timestamps, tool calls, and pass/fail status.

PR workflow with review gates

Agents open PRs and post drafts for review. You authorize merges; nothing ships without command approval.

Shared codebase memory

Branch strategy, lint rules, and reviewer preferences persist across agents and sessions.

GitHub-native execution

Real commits, real branches, real CI — connected through OpenClaw skills and your private runtime.

[05] Governance

You're in charge.

Team members

200 included

Workspaces

100 included

Unlimited agents and chats
Adaptive memory and shared workflows
Skills, integrations, and browser automation
Private VPC or on-prem deployment
SSO, custom domains, and agreements
Priority support with SLA

Talk to sales

See full rate card and FAQ

[07] Message from the founder

Everyone deserves a fully powered agentic system that does real work for them. That power should not sit only in the hands of big corporations — so we built Trooper for you, and made it free for anyone to use.

Vaibhav

Founder, Trooper

@absurdfounder

[08] Intel Brief

Intel brief.

Missing intel? Transmit your question to vaibhav@trooper.so.

Trooper for CodingUse Codex & Claude Code together.

Multi-agent coding ops

Multi-agent harness

Unified task board

BYOA subscriptions

Build agents that run your playbooks.

How this unit runson Trooper.

Your subscriptions. Our harness.

Fix invoice parser null rows

Diffs, CI logs, and PR bundles on Canvas.

Branch rules and reviewer prefs persist across agents.

Parser hotfix bundle on Canvas.

How Trooper runs your code unit

Parallel patch missions

PR workflow with review gates

Shared codebase memory

GitHub-native execution

You're in charge.

Simple pricing, by deployment.Pay for the plan you need.

Local Install

Solo Cloud

Trooper Cloud

Enterprise

Local Install

Solo Cloud

Trooper Cloud

Enterprise

Intel brief.

Try Trooper now.

Trooper for Coding
Use Codex & Claude Code together.