MCP Server

Connect Claude Code, Cursor, Codex, Continue, or Linear Agent to QA.tech so your AI assistant can list test cases, start runs, and read results.

The QA.tech MCP server exposes your test cases, runs, issues, knowledge base, and agent rules to any AI client that speaks the Model Context Protocol. Once connected, your assistant can answer questions like “what runs failed today?” or “rerun the failed cases from run UkxK” without you leaving the chat.

Accessing the MCP

The MCP is available at:

https://api.qa.tech/v1/mcp

Authentication is handled via OAuth.

• On first connect you sign in to QA.tech and pick an organization (and optionally specific projects).
• You can only access data you have permission to view in that organization, limited to the projects you approved.

API keys still work for CI and clients that don’t speak OAuth yet. Open the dashboard at Organization Settings → MCP Server for copy-paste client snippets.

Pick Your Client

Most modern clients have a one-click or one-line install. Clients that support MCP OAuth (Claude Code, Cursor, Codex, Linear Agent) only need the URL above. API keys still work for CI and clients that don’t do OAuth yet.

Claude Code

Run this in any project where you use Claude Code. Requires the Claude Code CLI.

claude mcp add --transport http qatech \
  'https://api.qa.tech/v1/mcp'

Then run /mcp in Claude Code (or claude mcp login qatech) and complete the browser sign-in. Claude Code reloads MCP servers automatically.

Prefer a static key instead? Add --header 'Authorization: Bearer <API_KEY>' from Organization Settings → API Keys.

Cursor

Go to Settings → Cursor Settings → Tools & MCPs → New MCP server. Cursor opens ~/.cursor/mcp.json. Paste this in — Cursor will open a browser window for you to sign in to QA.tech and approve access:

{
  "mcpServers": {
    "qatech": {
      "type": "http",
      "url": "https://api.qa.tech/v1/mcp"
    }
  }
}

Prefer a static key instead? Add an Authorization header with Bearer <API_KEY> from Organization Settings → API Keys.

Codex

Run this command to add the MCP server to Codex. Requires the Codex CLI. Codex starts the OAuth browser flow when the server supports it; you can also run codex mcp login qatech.

codex mcp add qatech \
  --url 'https://api.qa.tech/v1/mcp'

Prefer a static key instead? Add --header 'Authorization: Bearer <API_KEY>'.

First time using an MCP server in Codex? Add experimental_use_rmcp_client = true under [features] in ~/.codex/config.toml.

Linear

Add a custom MCP server in Linear Agent with URL https://api.qa.tech/v1/mcp (OAuth, or a Bearer API key if Linear cannot complete OAuth).

Full workspace enablement, Agent personalization steps, and how this differs from exporting issues are on Linear. Linear’s own UI is in Connect MCP Servers to Linear Agent.

Continue (VS Code)

Continue uses a static API key. Add this snippet to ~/.continue/config.yaml, then reload the Continue extension.

mcpServers:
  - name: qatech
    type: streamable-http
    url: https://api.qa.tech/v1/mcp
    requestOptions:
      headers:
        Authorization: Bearer <API_KEY>

Or Use the QA.tech CLI

The QA.tech CLI can write the config for any supported client in one command. Useful if you switch machines often or want the same setup scripted.

Don’t have the QA.tech CLI? Install it with npm install -g @qadottech/cli.

For Cursor and Claude Desktop, the default is a URL-only OAuth entry (no API key):

qatech mcp configure --client cursor
qatech mcp configure --client claude-desktop

Clients that still need a Bearer token (Continue, Goose) require an API key first:

qatech configure -k <your-api-key>
qatech mcp configure --client continue

Pass --api-key to force a static Bearer entry for Cursor/Claude Desktop. Add --print to dump the snippet without writing it. See qatech mcp configure --help for all flags.

What Your Assistant Can Do

The server exposes 43 tools: 24 that need read scope and 19 that need write scope. Tools your credential lacks the scope for are hidden from tools/list, so a read-only API key never sees the write tools. OAuth grants include both scopes.

MCP clients that group tools by risk (Claude Desktop Tool Permissions, ChatGPT) read the protocol annotations (readOnlyHint / destructiveHint) derived from those scopes, not the scopes themselves.

Most tools are project-scoped and take a projectShortId. Call list_projects first to find it, or use a project-scoped API key, which falls back to its own project.

Runs and results

Tool	Scope	What it does
list_runs	read	Lists runs newest first. Returns the 20 most recent unless you pass a since or until window.
get_run	read	Fetches a run by short ID. Pass testCases: "all" or "failed" for nested results, and issues to include them.
get_run_issues	read	Lists the issues found during a run, deduplicated across its test cases and filterable by issueType.
get_run_test_case	read	Fetches one test’s execution in a run: goal, evaluator reasoning, result, error classification, and media links.
get_run_test_case_trace	read	Returns the agent’s step-by-step trace for a run test case, with reasoning, tool calls, and screenshots.
get_test_case_history	read	Returns recent pass/fail history and durations for a test case, to tell flaky apart from consistently failing.
start_run_with_test_cases	write	Starts a run for an explicit set of test cases.
start_run_with_test_plan	write	Starts a run from a saved test plan.
rerun_run	write	Reruns a previous run. Pass failedOnly: true, or projectTestCaseIds to rerun a subset.

Test cases

Tool	Scope	What it does
list_test_cases	read	Lists test cases, filterable by application, labels, enabled state, or scenario.
get_test_case	read	Fetches a test case’s authored configuration: goal, expected result, steps, dependencies, labels, and scenario.
create_test_case	write	Creates a test case in draft. A burn-in run starts automatically.
update_test_case	write	Updates a test case’s authored configuration. Only the fields you pass change.

Issues

Tool	Scope	What it does
list_issues	read	Lists the issues found in the project, newest first. Defaults to ACTIVE; filter by severity or status.
get_issue	read	Fetches one issue by short ID, with its tags and the recent run test cases where it was detected.

Projects, applications, and environments

Tool	Scope	What it does
list_projects	read	Lists the projects the credential can reach. Pass a returned shortId as projectShortId.
create_project	write	Creates a project. Requires an organization-scoped credential.
list_applications	read	Lists the applications under test in a project.
list_application_environments	read	Lists the environments configured for one application.

Mobile builds

Tool	Scope	What it does
application_build_upload_url	write	Returns a presigned upload URL and a build token for an .apk or .ipa file.
create_application_build	write	Registers a build uploaded with that token and returns an ID you can attach to a run.

Both tools back mobile app testing. Upload the file to the presigned URL yourself with an HTTP PUT between the two calls.

Knowledge base

Tool	Scope	What it does
list_knowledge_items	read	Lists knowledge items (notes, crawled links, PDFs, agent memories, icons) without content.
get_knowledge_item	read	Fetches one knowledge item including its full content.
create_knowledge_item	write	Adds a text note, a URL to crawl, or an agent memory, and indexes it. PDF and icon items are managed in the dashboard.
update_knowledge_item	write	Updates a text, link, or memory item. Changing the content or URL re-indexes it.
delete_knowledge_item	write	Deletes a knowledge item and removes it from the index.

Rules

Tool	Scope	What it does
list_rules	read	Lists the project’s rules, filterable by the agent that consumes them.
create_rule	write	Creates a standing instruction for a QA.tech agent, optionally scoped to applications, labels, scenarios, or URL paths.
update_rule	write	Updates a rule’s title, content, filters, or sort order.
delete_rule	write	Deletes a rule so it is no longer injected into the agent’s context.

Features

Tool	Scope	What it does
list_features	read	Lists the project’s features: groupings of related UI actions discovered while crawling your product.
get_feature	read	Fetches one feature with the pages it spans and the actions that belong to it.
search_features	read	Searches the features by meaning rather than string match. Accepts 1-10 phrasings of the same concept.

Chat

Tool	Scope	What it does
create_chat	write	Starts an AI chat conversation and sends its first message.
append_chat_message	write	Sends a message to an existing conversation.
get_chat_conversation	read	Returns conversation metadata and recent messages, newest first.

Release checks

Tool	Scope	What it does
start_release_check	write	Starts a release check: a regression suite run, an exploratory review of your changes, or both, in one report.
get_release_check	read	Fetches the compiled report. Poll until status is completed.

Remote tunnels

Tool	Scope	What it does
create_remote_tunnel	write	Creates a tunnel that exposes a local port to QA.tech for testing.
list_remote_tunnels	read	Lists the project’s remote tunnels.
get_remote_tunnel_status	read	Returns the live health status of one tunnel.
delete_remote_tunnel	write	Tears down a tunnel and its DNS records.

Feedback

Tool	Scope	What it does
report_feedback	read	Sends feedback about the MCP tools to the QA.tech team when a tool is missing, unclear, or returns unexpected output.

Example prompts

“Which of my test cases failed in the last run?”

“Rerun run UkxK, but only the failed cases.”

“Show me the agent’s trace for the failing checkout test and tell me where it went wrong.”

“List the critical issues QA.tech found this week, then help me fix the first one.”

“Create a test case titled ‘Checkout with expired card’ for the frontend app.”

“Add our new API error codes to the knowledge base so the agents stop reporting them as bugs.”

“Which features cover inviting a teammate?”

Authentication

Two options:

OAuth (recommended for interactive clients)

Claude Code, Cursor, Codex, Linear Agent, and other MCP clients that support OAuth discover QA.tech’s authorization server automatically. On first connect you sign in with your QA.tech account, pick an organization, and approve access. Tokens refresh in the background. OAuth grants include both read and write, so every tool is available for the organization and projects you approved.

API key

Same Bearer tokens as the REST API. Useful for CI and clients that don’t speak OAuth yet.

• Where to find your key: Organization Settings → API Keys in the dashboard.
• Scopes: read and write are independent. A key with read sees the 24 read tools, a key with write sees the 19 write tools, and a key with both sees all 43.
• Project binding: Keys may be org-scoped or project-scoped. Org-scoped keys need a projectShortId on project-scoped tool calls; project-scoped keys fall back to their own project.

Treat API keys as confidential credentials. Anyone with the token can read your test data and, with a write key, create and rerun tests. Rotate keys from the dashboard if one leaks.

Endpoint Details

Field	Value
URL	https://api.qa.tech/v1/mcp
Transport	Streamable HTTP (JSON-RPC 2.0), stateless
Methods supported	initialize, tools/list, tools/call
Auth	OAuth (MCP discovery) or Authorization: Bearer <API_KEY>
Protocol version	2025-06-18
Server name	qatech

Batch JSON-RPC requests and GET upgrades are not supported. Each call is a single POST with a JSON body.

Smoke test with curl

curl -sS -X POST 'https://api.qa.tech/v1/mcp' \
  -H 'Authorization: Bearer <API_KEY>' \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

A successful response lists the tools your key has access to.

Troubleshooting

Symptom	Likely cause
Client lists no QA.tech tools	Restart the client after editing config. Verify the JSON or YAML parses.
Browser never opens for OAuth	Confirm the client supports MCP OAuth and the config has no Authorization header (URL only). Retry from the client’s MCP login / /mcp UI.
401 or 403 from the server	OAuth consent not completed, wrong org, or API key missing/revoked. Re-auth or re-copy the key.
Some tools missing	API key lacks write scope. Generate a write key, or use OAuth (includes write).
Claude ofid_… after sign-in	Consent succeeded; Claude’s follow-up MCP initialize failed. Confirm the URL is https://api.qa.tech/v1/mcp. Retry, or use an API key.
405 Method Not Allowed	The client sent a GET. QA.tech MCP is POST-only and stateless.
qatech mcp configure says no API key	Continue/Goose need a key: run qatech configure -k <key> or pass --api-key. Cursor/Claude Desktop default to OAuth and do not need one.

For anything else, contact support.