# ExpLint

> ExpLint lints A/B test specs against the rules a team writes in plain English, reporting each rule as an error, a warning, or a pass together with the reason behind the call, and an assistant reaches it through a remote MCP server while a script or an agent that sends HTTP gets the same work over a REST API.

An agent working with ExpLint usually starts by connecting the MCP server and reading the documentation it returns, because that one document covers how to sign in, how projects hold the rules a lint runs against, and what every tool and endpoint takes and returns. Projects are named inside the organization the credential belongs to, so no call names an organization, and a lint sends a spec as markdown, HTML, or plain text and answers with the rules it broke, the rules that only warn, and the rules it passed, each with the chance the model gave the spec of satisfying that rule.

Follow the links below for the detail behind each surface, and fetch only the file a task needs so the rest stays out of context.

## Documentation

- [ExpLint API documentation](https://explint.com/docs): The developer documentation as a web page, covering installation and sign-in for the MCP server, every MCP tool, the REST endpoints beside them, and what each refusal status means.
- [ExpLint documentation as markdown](https://explint.com/docs.md): The full ExpLint documentation in one markdown file with the same content the documentation page carries.
- [ExpLint OpenAPI document](https://explint.com/openapi.json): The ExpLint REST API described as an OpenAPI document in JSON, with every path, argument, answer shape, and the OAuth scope each operation names.
- [ExpLint OpenAPI document as YAML](https://explint.com/api/openapi.yaml): The same ExpLint OpenAPI document in YAML, also served in JSON at https://explint.com/api/openapi.json.
- [ExpLint REST API OAuth metadata](https://explint.com/.well-known/oauth-protected-resource/api): RFC 9728 protected-resource metadata for the ExpLint REST API, naming its authorization server and scopes_supported.
- [ExpLint MCP server](https://explint.com/mcp): The remote ExpLint MCP endpoint an assistant connects to, with OAuth sign-in through the browser and an API key in an Authorization header for clients that cannot run that flow.
- [ExpLint MCP OAuth metadata](https://explint.com/.well-known/oauth-protected-resource/mcp): RFC 9728 protected-resource metadata for the ExpLint MCP server, naming its authorization server and scopes_supported.

## Guides

- [Check Jira experiment tickets before the sprint](https://explint.com/guides/jira): Read Jira experiment tickets with Atlassian's Rovo MCP server, lint each description and acceptance criteria, then comment the findings on the ticket.
- [Lint Linear issues while the idea is still a draft](https://explint.com/guides/linear): Read Linear issues from a project or label, lint each description as one text, and post the findings as a comment once you approve the wording.
- [Catch vague Trello cards before they reach Ready](https://explint.com/guides/trello): Read every card in an Ideas list, group the gaps the lint finds, and write the findings and a rewrite back onto the card you pick.
- [Review GitHub experiment issues in the issue thread](https://explint.com/guides/github): Read experiment issues and their Project status field, lint each body, and leave the findings and a suggested rewrite in the issue thread.
- [Audit Asana experiment tasks with their subtasks](https://explint.com/guides/asana): Read Asana experiment tasks with their subtasks and custom fields, lint the combined text, and comment on the task where planning already happens.
- [Check Notion experiment briefs against their status](https://explint.com/guides/notion): Search Notion for briefs in a chosen status, lint each page with its properties, and leave the findings as a comment on the brief.
- [Lint ClickUp tasks from the summary and description](https://explint.com/guides/clickup): Read ClickUp experiment tasks from their summary and description, lint what comes back, and post the findings as a comment on the task.
- [Check monday.com experiment items column by column](https://explint.com/guides/monday): Read a monday.com board's items with their columns and subitems, lint each one, and post the findings as an update on the item.
- [Audit GitLab work items before the sprint starts](https://explint.com/guides/gitlab): List opened GitLab work items by label, lint each title and description, and add the findings and a rewrite as a note when you approve.
- [Catch Azure Boards work items missing a metric](https://explint.com/guides/azure-boards): Read Azure Boards work items by state, lint each title, description, and tags, and comment the findings on the item you pick.
- [Check a GrowthBook draft before it goes live](https://explint.com/guides/growthbook): Read a GrowthBook draft through its MCP server, lint the hypothesis, variations, and metrics, and map each finding to the field that fixes it.
- [Review Statsig allocation and metrics before launch](https://explint.com/guides/statsig): Read a Statsig experiment's groups, allocation, and metrics, lint them as one text, and name the setting behind every finding.
- [Lint an Eppo plan from the API or a paste](https://explint.com/guides/eppo): Eppo publishes no MCP server, so bring the plan by API or paste; the lint returns the fields a launch review would ask about.
- [Audit Optimizely variations, audience, and split](https://explint.com/guides/optimizely): Read an Optimizely draft's variations, audience, and split, lint them as one text, and rank the findings by what you would fix first.
- [Check a VWO campaign's goals before it takes traffic](https://explint.com/guides/vwo): Read a VWO campaign's hypothesis, variations, and goals, then lint them before the campaign takes a single visitor.
- [Review an Adobe Target activity before activation](https://explint.com/guides/adobe-target): Read an Adobe Target activity's audience, experiences, and reporting metrics, and give the approver the findings before activation.
- [Check a PostHog experiment's sample size and metrics](https://explint.com/guides/posthog): Read a PostHog experiment with its sample size estimate, lint metrics and variants, and separate writing fixes from setup fixes.
- [Audit AB Tasty feature tests and web campaigns](https://explint.com/guides/ab-tasty): Check an AB Tasty feature test through its MCP server, or a web campaign by API or paste, and tie every finding to a setup step.
- [Check a Kameleoon experiment through the Automation API](https://explint.com/guides/kameleoon): Read a live Kameleoon experiment through the Automation API, lint variations, audience, and goals, and ask its docs server for context.
- [Audit LaunchDarkly experiments and their treatments](https://explint.com/guides/launchdarkly): Read a LaunchDarkly experiment with treatments, allocation, and targeting, then lint them and say where a guardrail belongs.
- [Check Amplitude variants and metrics before launch](https://explint.com/guides/amplitude): Read an Amplitude experiment's variants, allocation, metrics, and dates, then lint them and flag metrics that arrive as bare ids.
- [Review Harness FME experiments and their flags](https://explint.com/guides/harness-fme): Read a Harness FME experiment beside the flag carrying its allocation, lint the pair, and name the Harness field behind each finding.
- [Audit Convert experiences and their traffic split](https://explint.com/guides/convert): Read a Convert experience's objective, audience, and traffic distribution, lint them, and stage any change for your review before it saves.
- [Check a pasted Dynamic Yield campaign plan](https://explint.com/guides/dynamic-yield): Paste a Dynamic Yield campaign plan or report export, lint it as one text, and get back the headings that are empty.
- [Lint a Firebase A/B test from the Remote Config API](https://explint.com/guides/firebase): Read a Firebase A/B test through the Remote Config experiments API, lint what it returns, and name what only the console knows.
- [Check SiteSpect campaign metrics and sample size](https://explint.com/guides/sitespect): Read a SiteSpect campaign through the Admin API, lint its metrics and sample size, and tie each finding to a campaign section.
- [Audit Unleash experiment flags and their variants](https://explint.com/guides/unleash): Read an Unleash experiment flag's rollout and variants, lint what the flag carries, and list the plan's parts that need your words.
- [Check a Flagsmith experiment's split before launch](https://explint.com/guides/flagsmith): Read a Flagsmith experiment's hypothesis, split, and primary metric, then lint the setup before the flag takes traffic.

## Optional

- [ExpLint homepage](https://explint.com/): The product overview with a working linter on the page, the plans, and the answers to the questions people ask before installing anything.