Open Source · MIT Licensed

Project management
for AI agents.

Tasks, issues, and persistent memory for Claude Code, Cursor, Windsurf, and any MCP-compatible client. Your agents remember decisions, track work, and start every session with full context.

Free cloud tier available. Self-hosted coming soon.

claude-code

31 tools · 7 categories · works today

Works with any MCP-compatible AI agent

Claude Code
Cursor
Windsurf
VS Code
Any MCP Client

31 tools. Zero forgotten context.

Everything your AI agents need to manage work across sessions -- tasks, issues, memories, projects, cycles, and config -- all through the Model Context Protocol.

Tasks & Issues

Full-featured project management, built for AI.

Your AI agents create, update, and complete tasks the same way a human engineer would. Custom workflows, priorities, labels, cycles, and estimates -- all through MCP tool calls.

  • Custom workflow statuses (backlog, in progress, review, done)
  • Priority levels, severity, labels, and estimate points
  • Sprint cycles with start/end dates and progress tracking
  • Assignees for human-AI collaboration on shared boards
payments-apiSprint 3
5 tasks
Add idempotency keys to charge endpoint
CL
Implement webhook signature verification
AI
Add Stripe test mode toggle to settings
Implement webhook retry logic
Write integration tests for refund flow
Persistent Memory

Your AI remembers everything.

Every architectural decision, debugging insight, and team preference is stored and recalled automatically. No more re-explaining your codebase every session.

  • Four memory types: facts, decisions, preferences, learnings
  • Tag-based organization and full-text search
  • Memories scoped per project, recalled per session
  • Semantic search coming soon for intelligent recall
Memoriespayments-api
12 total

Auth uses JWT with RS256 algorithm and 24h token expiry

fact#auth#security3d ago

Exponential backoff with jitter for webhook retries, max 5 attempts

decision#webhooks#reliability1d ago

Always use Zod schemas for API request validation, not manual checks

preference#api#validation5d ago

Stripe webhook events can arrive out of order - must handle idempotently

learning#stripe#gotcha2d ago
Project Context

Full context, every session.

One tool call at the start of every session loads pending tasks, open issues, relevant memories, project config, active cycle, and your custom AI persona. Zero manual setup.

  • Active project, cycle, and config in a single call
  • Pending tasks and open issues loaded automatically
  • Recent memories surface relevant past decisions
  • Custom AI persona shapes agent behavior per project
get_context
loaded in 120ms
Projectpayments-api
Active CycleSprint 3 (4 days left)
Pending Tasks3 tasks across 2 labels
Open Issues2 issues (1 critical)
Memories Loaded12 memories (4 decisions, 3 facts, 3 learnings, 2 prefs)
AI PersonaSenior backend engineer, concise, test-driven

See it in action

Tasks, issues, and memories working together across sessions — your AI never starts from scratch.

Session 1Monday 2:30 PM
Create a task to implement Stripe webhook handling with high priority
Created task PAYMENTS-7: Implement Stripe webhook handling [High Priority]
Task created
Remember that we decided to use idempotency keys for webhook deduplication
Saved as a decision memory in payments-api.
Memory saved
New Session
Session 2Wednesday 10:15 AM
What's on my plate for the payments project?
Loading context... You have 3 pending tasks including PAYMENTS-7 (Stripe webhooks). I also found a relevant memory: you decided to use idempotency keys for webhook deduplication.
Context loaded1 memory recalled
Found a bug — webhook events are being processed twice
Created issue PAYMENTS-12: Duplicate webhook event processing [Critical]. Based on your earlier decision to use idempotency keys, this might be a race condition in the queue consumer.
Issue createdMemory applied
60 seconds to setup

60 seconds to an AI that never forgets

Three steps. No complex setup. Start with our cloud — self-hosted mode is coming soon.

01

Sign Up

Create your free account at dodev.ai. One click with Google or GitHub — no credit card, no setup wizard.

Get Started FreeGoogle or GitHub — takes 5 seconds
02

Connect Your Agent

Add the MCP server config to Claude Code, Cursor, or any MCP-compatible client.

mcp.json
{
  "dodev": {
    "type": "http",
    "url": "https://mcp.dodev.ai/mcp"
  }
}

Note: Your agent will be prompted to authenticate on first use.

03

Start Building

Your AI now has persistent tasks, issues, and memories across every session. Context that carries forward, always.

Terminal

$ claude

> "What did we decide about the auth strategy?"

Found 3 memories for project backend-api

JWT with RS256, 24h expiry, refresh tokens via httpOnly cookies

> "Create a task to add rate limiting to the auth endpoints"

Task BACK-42 created: Add rate limiting to auth endpoints

Simple, transparent pricing

Free to start, upgrade when you need more. Self-host coming soon.

Hobby

$0/month

Get started with generous free limits

  • 1 project
  • 1 agent connection
  • 100 tasks & issues
  • 30 memories
  • 31 MCP tools
  • Web dashboard
  • Cloud MCP server
  • Community support
Join Waitlist

Personal

$10/month

More projects and capacity for solo builders

  • 3 projects
  • Up to 3 agents
  • 500 tasks & issues
  • 200 memories
  • 31 MCP tools
  • Web dashboard
  • Cloud MCP server
  • Priority support
Join Waitlist
Popular

Pro

$20/month

Unlimited everything for serious builders

  • Unlimited projects
  • Up to 10 agents
  • Unlimited tasks & issues
  • Unlimited memories
  • Semantic memory search
  • Advanced analytics
  • Priority support
Join Waitlist
Coming Soon

Self-Hosted

Freeforever

Your infrastructure, your data. Run on your own servers with Docker or npx.

  • Unlimited everything
  • Docker & npx install
  • No account required
  • Full data ownership
  • Community support
Roadmap

What's coming next

We're building in public. Here's what's on our radar — and we ship fast.

Coming Soon

Self-Hosted Mode

Run on your own infrastructure with Docker. Your data, your servers. No account required.

Coming Soon

Documentation

Comprehensive docs site with guides, API reference, and integration tutorials for every MCP client.

Coming Soon

Semantic Memory Search

Vector embeddings for intelligent memory recall. Your AI finds relevant context even with different wording.

Coming Soon

Team Collaboration

Share projects, memories, and context across your entire team of AI agents. Collective intelligence.

Want to shape what we build next?

Star us on GitHub, open an issue, or join the conversation.

Stop re-explaining yourself
to your AI.

Give your AI persistent memory. Join the developers who are building with continuity, not repetition.

Free forever for self-hosted. No credit card required.