TestNeo
TestNeo Docs Developer documentation
Documentation

Build with TestNeo, your way.

Release Intelligence from your IDE, plus SDK, CLI, MCP, NLP Web & API, and k6 Load Insights — one platform for PR risk, tests, load evidence, and ship confidence.

Start here

Release Intelligence

Validate PRs from Cursor, VS Code, or GitHub — PASS / WARN / BLOCK, execute impacted tests on PR, Engineering Memory, release gate, and outcome learning. Includes GitHub PAT setup and a full UI walkthrough (index → Memory → Release confidence).

  • GitHub PAT scopes (classic + fine-grained) before connect
  • UI walkthrough: connect → index → memory → PR → release confidence
  • Full business flow: ingest → patterns → validate → gate → outcomes
  • Copy-paste prompts EM1–EM9, C17–C25 · GitHub @TestNeo validate
  • Outcome learning + read-only calibration
Release Intelligence guide →
AI agents

AI Agent Verification

Prove agent runs with Claimed → Observed → Verified. Install testneo-agent-sdk, ingest a summary, open Agent Verification for PASS / WARN / BLOCK — tools, RAG, memory, and live API checks.

  • 15-min SDK quickstart (CrewAI / LangGraph / AutoGen)
  • Eval lenses: Tools / RAG / Memory
  • Two lanes: product API vs research / content crews
  • API NLP from observed HTTP — no invented UI tests
Agent verification guide →
Integrate

How to integrate

Install testneo-agent-sdk (recommended), or push via curl / MCP / n8n. Ingest contract + mock refund probe.

  • Push the run, then run verification
  • Local demo with a mock refund API
  • Works with n8n + MCP
Integration guide →
Integrations

n8n templates

Download importable HTTP workflows for post-agent verification, AgentRunSummary ingest, and release outcomes — same REST as MCP. Not a Marketplace node.

  • Download Template 1–3 JSON from the docs page
  • Import into n8n → set API key → run
  • Sample webhook + outcome payloads included
  • No repo clone required for customers
n8n templates guide →
Load testing

k6 Load Export & Insights

Export the same API NLP suite as a Grafana k6 kit, run load on your laptop or CI, upload summaries back for p95, error rate, and verdict history.

  • Step-by-step from Test Manager → kit ZIP
  • Local / Docker run · tweak VUs & BASE_URL
  • Import testneo-import.json → Load Insights
  • Smoke / load / ramp / spike / soak profiles
k6 how-to guide →
For developers

Playwright AI SDK

Write Playwright tests in plain English, keep full Playwright control, and publish runs to TestNeo from your CI.

  • ai.run, ai.step, ai.api in one file
  • Hybrid mode: NLP + native page.click()
  • Strict / balanced / adaptive modes for CI
  • Guest mode — try without an API key
Read SDK docs →
New · Claude

Claude Desktop + MCP

Configure Claude Desktop in ~10 minutes. Login runs, release readiness, PR Release Brief, and Engineering Memory — ordered demo flows with copy-paste prompts.

  • macOS / Windows / Linux config paths
  • Read-only smoke → enable writes → run login
  • Demo flows: readiness → PR Brief → explain/fix
  • Prompt cookbook: discover, execute, generate, PR & memory
  • Security FAQ for customers (what Claude sees)
Claude Desktop guide →
For your whole team

MCP AI Agent

Add TestNeo to Cursor, Claude Desktop, VS Code, or any MCP client. Generate, run, watch, and triage from one thread.

  • One flow: generate → run → watch → triage
  • Plain-language answers for every role
  • Writes need your confirmation (on by default)
  • Failures: cause + rerun plan in chat
  • 83 MCP tools · AV1–AV5 agent prompts · EM1–EM9 · execute on PR
Read MCP docs →
For product & GTM

VS Code Extension

What TestNeo supports in VS Code today, how extension + MCP workflows fit together, and how teams move from code change to release confidence.

  • Source code → impact → validation workflows
  • Planning-only vs executed stage explained clearly
  • Positioning language for customers and field teams
  • Architecture map: Extension + MCP + Playwright + backend
Read VS Code docs →
Terminal & CI · v0.2.1

TestNeo CLI

Release assurance in your shell — run .testneo tests, validate PRs, Engineering Memory, readiness, and Ask. Smart local-agent routing.

  • run / try / --pick for execution
  • Local agent when online · --cloud / agent status
  • pr validate · memory · readiness
  • ask — AI Assistant in the terminal
Read CLI docs →
Test Manager

NLP Web & API Testing

Write browser and REST tests in plain English — Steps that run, Objective goals that Generate steps, and data-driven CSV/JSON loops.

  • Objective vs Steps — Run always executes Steps
  • Web UI + API NLP recipes (FakeStore, SauceDemo)
  • Data-driven: Data: vs Use data from
  • Good vs broken Objective goals for CSV loops
NLP testing guide →

Quick links

NLP Web & API guide Data-driven CSV / Use data from Objective goals → Generate steps k6 Load Insights Export kit → local run → upload Claude Desktop demo flows Claude PR / readiness prompts Claude Desktop + MCP CLI Quickstart SDK Quickstart Release Intelligence Engineering Memory flow Workflow 7 — PR validate Patterns & trend GitHub PAT & permissions UI: PR → Memory → Release GitHub @TestNeo validate Code index from GitHub Hands-on demo (15 min) Outcome learning Workflow 8 — Engineering Memory Workflow 10 — Ship loop MCP Quickstart VS Code Extension Guide MCP Tool Reference SDK API Reference MCP Workflows MCP Customer playbook MCP Prompt Packs EM prompts (EM1–EM9) MCP Web AI Assistant Try the Playground