ThinkNCollab Documentation
Everything you need to build, test, automate, and scale with ThinkNCollab. Explore our CLI tooling, Model Context Protocol (MCP) servers, automated task judge, zero-docker sandboxes, and REST APIs.
thinknsh CLI
Command-line suite for developers to run tasks, spin up sandboxes, trigger judge verification, and sync git branches.
$npm i -g thinknsh$thinknsh login$thinknsh tasks listModel Context Protocol (MCP)
Connect Claude Desktop, Cursor, ChatGPT, and custom LLM agents directly to ThinkNCollab workspaces via JSON-RPC 2.0 & SSE.
{ "mcpServers": { "thinkncollab": { "command": "npx", "args": ["-y", "@thinkncollab/mcp-server"] } }}Task Verification & Judge Engine
Objective verification engine that grades code changes against structured Gherkin criteria, test suites, and audit trails.
$thinknsh tasks submit 6a318170496c7b00a7f74260[Judge] Executing test assertions...✓ HTTP GET /health -> 200 OK✓ Verdict: PASSED (Score: 100%)Zero-Docker MicroVM Sandboxes
Instant Linux sandbox environments operating without Docker daemon overhead. Test PRs and experiment in complete isolation.
$thinknsh sandbox create --runtime node20⚡ Booting MicroVM sandbox (420ms)...✓ Interactive bash session ready.Marketplace & Partner Platform
Build, distribute, and monetize third-party integrations, webhook consumers, and automated bots across ThinkNCollab rooms.
$curl -X POST https://thinkncollab.com/api/marketplace/apps \ -H 'Authorization: Bearer tn_sec_...' \ -d '{"name": "DeployBot", "scopes": ["tasks:write", "rooms:read"]}'Agent OS REST API & OpenAPI 3.0
Standardized HTTP REST API covering rooms, kanban boards, tasks, microVM sandboxes, and autonomous AI connectors.
$curl https://thinkncollab.com/.well-known/openapi.json# OpenAPI 3.0 JSON schema$curl https://thinkncollab.com/.well-known/ai-plugin.jsonEngineering Methodologies & Playbooks
Field-tested operational methodologies, DORA metrics, architectural decision records, and sprint retrospective standards.
# DORA Metrics Standard in ThinkNCollabDeployment Frequency: On-Demand (Multiple/Day)Lead Time for Changes: < 1 HourChange Failure Rate: < 5%Mean Time to Recovery: < 15 MinsCI/CD Pipelines & Load Testing
Native continuous integration pipelines and high-concurrency API stress testing engines running right inside your rooms.
$thinknsh loadtest run --target https://api.mysite.com \ --concurrency 50 --requests 1000[GeekLoad] 99th percentile: 38ms | 0 errorsConnect your agent or CLI in under 60 seconds
ThinkNCollab bridges local developer workstations, autonomous AI copilots, and cloud environments seamlessly. Everything from branch synchronization to automated test verification works out of the box.
// 1. Install thinknsh globally
$ npm install -g thinknsh
// 2. Authenticate securely with one click
$ thinknsh login
// 3. View room tasks & checkout your active sprint
$ thinknsh tasks list
// 4. Submit code for automated judge verification
$ thinknsh tasks complete <taskId>