Tesults Mcp

io.github.tesultsPublished by GitHub account tesultsv1.1.0Updated Oct 25, 2026

Ask your AI agent why your tests are failing. Instant analysis on failures and regressions.

VerifiedSTDIODesktop onlyDeveloper ToolsData & Analytics

Overview

AI-generated overview

Connects an AI assistant to Tesults test results so it can list targets, inspect runs, and explain failures, regressions and flaky tests.

What you can do
  • List, create, update and delete targets (test jobs), and regenerate a target's API token
  • Retrieve test results for a target, including runs with pass/fail counts and individual test cases
  • Explain a specific run in plain language: summary, likely causes, non-passing count and failure patterns
  • Find flaky tests across recent runs and compare runs to surface new, resolved and continuing failures
  • Explain a single test case's stability, trend, patterns and recommendations from its recent history
When to use it
Useful when a team keeps test results in Tesults and wants an assistant to diagnose failing CI runs, spot regressions and flaky tests, or manage test targets without leaving the chat.
Requirements
Runs locally over stdio, typically via npx from the npm package. Requires a Tesults API token in the TESULTS_API_TOKEN environment variable, obtained from the Tesults project configuration. Optional TESULTS_READ_ONLY=true restricts it to reporting tools. Network access to Tesults is needed.
Before you install
The token grants the account's API permissions; keep it private. Creating, updating and deleting targets and regenerating tokens are write actions: deletion is permanent and removes all test results, and regenerating a token immediately breaks existing CI/CD pipelines using the old one. Read-only mode is a restriction in this server, not a reduction of the token's permissions, and insight tools may still generate and cache AI analyses on the server.

Installation

In SourceWeft

  1. Open Tesults Mcp in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

tesults-mcp

Tesults MCP server. Connect AI agents and MCP-compatible tools to your Tesults test results, insights, and targets.

Documentation

Full setup and usage documentation: https://www.tesults.com/docs/api/mcp

Quick start

claude mcp add tesults-mcp npx tesults-mcp -e TESULTS_API_TOKEN=your-api-token

Reporting-only mode

Set TESULTS_READ_ONLY=true in the MCP server's environment alongside TESULTS_API_TOKEN. This exposes only the seven reporting and insight tools. Target creation, editing, deletion and token regeneration are not registered, and non-GET API requests are blocked locally.

Existing configurations remain unchanged when this setting is omitted or false. This is a restriction in this MCP server, not a reduction in the API token's permissions. Keep the token private. Insight tools can still generate and cache AI analyses on the server; they do not change your test results or targets.

Requests time out after 60 seconds and responses are limited to 10 MiB. For large result sets, request fewer runs or use cases=false for summaries. API failures are returned as tool errors, not successful answers.

These options apply to this source version and require an updated installation once released. They do not change an already installed published package.

Offline checks

Run node --test tests/hardening.test.js. Tests do not call Tesults or use a real token.

License

MIT

Source: README.md at commit 7798be8

Tools

11

tesults_get_targets

List all targets (test jobs) in the Tesults project. Returns target ids and names needed by other tools.

Parameterslimitcursor

tesults_create_target

Create a new target (test job) in the Tesults project. Returns the created target and its API token.

Parametersnamerequiredgroupdescription

tesults_update_target

Update an existing target. Any combination of name, description, or group can be updated.

Parametersnamegrouptargetrequireddescription

tesults_delete_target

Permanently delete a target and all its test results. This action cannot be undone.

Parameterstargetrequired

tesults_regenerate_token

Regenerate the API token for a target. The existing token will stop working immediately. Any CI/CD pipelines using the old token must be updated.

Parameterstargetrequired

tesults_get_results

Get test results for a target. Returns test runs with pass/fail counts and individual test cases.

Parametersrunbuildcaseslimitcursortargetrequiredanalysispersistent

tesults_explain_run

Get a plain-language explanation of a specific test run: summary, likely causes, non-passing count, and key failure patterns. Causes are hypotheses, not verified facts. Analysis covers at most 100 non-passing tests; check truncation metadata.

Parametersrunrequiredtargetrequired

tesults_get_flaky_tests

Find inconsistent interpreted results across at most 10 recent runs in the requested range, including pass/unknown transitions. Returns failure rate and classification; an empty list is limited to this window, not proof of long-term stability.

Parameterstofromtargetrequired

tesults_get_regressions

Compare the latest run with its immediately preceding run. Distinguishes new failures from continuing failures. Not a comparison against the last passing run or a historical baseline.

Parameterstargetrequired

tesults_what_changed

Compare a specific run against the previous run. Returns resolved failures, new failures, and continuing failures.

Parametersrunrequiredtargetrequired

tesults_explain_case

Get stability classification, trend, patterns, and recommendations from a test case's recent history. This is not scoped to a historical selected run. Interpret conclusions within the returned runs_analyzed count.

Parameterscaserequiredtargetrequired

Version history

2
  1. v1.1.0LatestOct 25, 2026
  2. v1.0.2Sep 16, 2026

Score

80% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts
  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listedRequired15 pts
  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts

    Declare a license in the repository or the registry entry.