Free tier — no credit card

Lint your tickets before agents touch them

Speclint scores every GitHub issue before your AI coding agent sees it. Bad specs ship broken code. Good specs ship in one pass.

Get API KeyRead the docs
Built for Product Owners and Scrum Masters who run refinement sessions

How It Works

Install once. Gate your issues. Ship faster.

1

File a GitHub Issue

Create an issue the way you normally would. Speclint works with your existing GitHub workflow — no new tools to learn.

2

Speclint Scores It

The GitHub Action fires on issues.opened. Speclint scores the spec 0-100, posts the refined version as a comment, and adds the agent_ready label when it passes.

3

Agent Ships It

Your AI coding agent (Cursor, Codex, Claude Code, Copilot) picks up the agent_ready spec and implements it. Good specs ship in one pass.

Refine Your Backlog

Paste your messy backlog items below and watch AI transform them into structured, sprint-ready stories.

Input Your Backlog
Free tier: up to 5 items per request

See the Transformation

Watch how AI turns vague todos into actionable, sprint-ready stories

Before: Messy Input
- fix bug with login
- new feature for dashboard
- update documentation
- performance improvements
- mobile responsive design
Problems with this backlog:
Vague titlesNo prioritiesNo acceptance criteriaNo estimates
After: Clean & Actionable

Fix Session Timeout Authentication Bug

Priority: HighEffort: M

Users are getting logged out mid-session due to token expiration mishandling, causing data loss and support tickets.

  • User stays logged in for full session duration (8 hours)
  • Token refresh happens silently without page reload
  • +1 more...
bugauthcritical
Needs clarification:
  • ?Confirm whether token refresh applies to all auth methods or OAuth only

Implement Dashboard Analytics Overview

Priority: MediumEffort: L

Users lack visibility into usage patterns and key metrics, reducing engagement and making it hard to demonstrate ROI.

  • Dashboard shows daily/weekly/monthly active users chart
  • Key metrics load in under 2 seconds
  • +1 more...
featureanalyticsux

Update REST API Documentation

Priority: MediumEffort: S

Outdated API docs are causing integration delays and increasing support burden from developer customers.

  • All endpoints documented with request/response examples
  • Authentication flow documented with code samples
  • +1 more...
docsdeveloper-experience
Now you have:
Clear titlesProblem statementsAcceptance criteriaSize estimates

Ready for import into Jira, Linear, GitHub Issues, or any project management tool

Built for AI-Native Development

Stop shipping broken code from bad specs. Speclint gates your issues before agents touch them.

Completeness Score

Every issue gets a 0-100 score. Gate your CI on ≥70. Agents only touch agent_ready specs.

GitHub Action

Install once. Fires on issues.opened. Posts refined spec as comment. Adds agent_ready label automatically.

Structured JSON Output

Not a chat response. Typed schema: title, problem, acceptanceCriteria[], estimate, tags[], completeness_score. Parse it in your pipeline.

Codebase Context

Pass your tech stack. Get ACs that reference your actual architecture, not generic boilerplate. Pro feature.

Fits Into Your Existing Pipeline

CI Gate
Block agents from picking up issues below your completeness threshold. Ship only what's ready.
GitHub Native
Posts refined spec as a comment. Adds agent_ready label. No new tool for your team to learn.
Any Agent Stack
Works with Cursor, Codex, Claude Code, Copilot — anything that reads GitHub issues.

Use It Your Way

Web UI, AI assistant, or automated pipeline — all hit the same API.

Web UI

Refine items manually in the browser.

speclint.aiTry free →
MCP Server

Works in Claude Desktop, Cursor, and any MCP-compatible client.

npx refine-backlog-mcpnpm →
CLI

Pipe into scripts, GitHub Actions, and CI pipelines. No LLM required.

npx refine-backlog-cli --file backlog.txtGitHub →
Agent-Native

Built to Be Called by AI

Refine Backlog isn't just a web app — it's an API-first service designed for agents, scripts, and pipelines. Every surface is machine-readable from day one.

llms.txt

LLM-readable manifest. AI agents discover what this API does before calling it.

GET https://speclint.ai/llms.txtView llms.txt →
OpenAPI Spec

Machine-readable API spec. Plug into any OpenAPI-compatible tool, agent, or code generator.

GET https://speclint.ai/openapi.yamlView spec →
REST API

Direct programmatic access. Call from any script, agent, GitHub Action, or CI pipeline with a license key.

POST https://speclint.ai/api/lintGet API key →

Need programmatic access? Get a license key →

Simple, Transparent Pricing

Start free with no credit card. Upgrade when your agents are shipping at scale.

Free
$0

Score your first issues

  • 5 items per request
  • 3 requests per day
  • completeness_score included
  • No credit card required

No signup required • Start immediately

Most Popular
Pro
$29/month

For solo devs running agents

  • Unlimited requests
  • 25 items per request
  • codebase_context (pass your tech stack)
  • Priority support

Instant access • Cancel anytime

Team
$79/month

For small firms automating at scale

  • Unlimited requests
  • 50 items per request
  • codebase_context included
  • Dependency mapping (/api/plan)
  • SLA

Enterprise features • Custom pricing

All plans include data privacy protection and secure processing.
No long-term contracts • Cancel anytime

Your Data Stays Yours

We built Refine Backlog with privacy as a default, not an afterthought.

✅

No storage.

We don't store your backlog content. Processed in memory, returned to your browser.

🔒

Encrypted.

All data encrypted in transit via TLS.

🚫

No training.

Your data is never used to train AI models. Anthropic's commercial API contractually prohibits it.

⚠️

Not for sensitive data.

Don't submit PII, health records, or credentials. This tool is for backlog items.

Read our Privacy Policy · Terms of Service

Frequently Asked Questions

Everything you need to know about Refine Backlog

Still have questions? Contact our support team