Taskachu is the shared workspace Claude, Cursor and Codex all read from and write to — your always-on co-founder.
Half-sentence in. Brief, decisions, and acceptance criteria out. Your project memory grows with every drop.
Tasks decompose, prioritise, and flow straight into your development workflow with full context attached.
Creatives, content, launch prep — drafted in your voice in parallel with the build, not after it.
Eight primitives, one continuous loop. Each surface is built around the same idea: keep your AI in context, and keep your hands on the keyboard.
Send tasks with full context directly into your development workflow. Brief, decisions and linked docs travel with the task — the first response is already on-spec.
Markdown docs with hierarchy and a BlockNote editor. Decisions, briefs, runbooks, brand notes — written once, retrieved by every agent you delegate to. Stop pasting from Notion every morning.
Ideas decompose into tasks automatically. The AI ranks impact vs. effort and surfaces what to ship today.
Every handoff to Claude carries the task, the linked docs, and the decisions that shaped it — so the first reply is already on-brief.
Workspace-scoped retrieval over every doc, card and decision. Ask Taskachu in-app or reach the same RAG from Claude Code / Cursor over MCP — no copy-paste.
Every completed task triggers your pipeline. Failures route back to the AI with the full trace — most fix themselves before you notice.
Creative briefs, content calendars, launch campaigns — drafted in your brand voice and queued to your channels.
API keys, environments and rotation, with a clear structure your AI can reason about without ever seeing the values.
Drop in your first idea — Taskachu turns it into a backlog and hands the top card to Claude in under a minute.
Five steps that keep going. You drop in a thought. Taskachu plans it, helps you build it, checks it works, then helps you tell the world. And then it asks what's next.
Take a walk. Read your kid a bedtime story. Sleep in.
The loop keeps running.
Last Tuesday I opened my own Backlog board, asked Claude Code to pick the top priorities, and handed each card off over MCP with full context attached. Thirty minutes later, the entire 120-task backlog had moved into Complete. Auth fixes. A pricing tweak. The PWA manifest. The very landing section you're reading.
No prompt re-typing. No tab-juggling. The loop just ran.
Open a workspace, plug Claude over MCP, ship your own backlog before lunch.
One workspace for backlog, docs and AI handoffs. You operate — the loop runs.
The backlog re-ranks itself. Today's top card is today's right move.
RAG + MCP let every agent read your docs, decisions and board. No paste-and-pray.
Solo is free forever. Upgrade to Team when you need more boards, teammates, RAG documents, and a co-founder that never sleeps. Start with a 7-day trial — no card required.
For founders who want to test the loop without committing.
Tag, drag, ship — all on the house.
Everything in Solo, plus the Pro-only modules.
Try free for 7 days — no card. Add one on day 7 for another 7 before auto-charge.
Drop in your first idea, watch it become a backlog, and hand the first task to Claude over MCP — all in one workspace.
Taskachu is an AI-native productivity app for solo founders. You manage projects on Kanban-style boards, and AI features (built into the app) help you decompose tasks, prioritise the backlog, run standups, and answer questions about your own workspace. Every AI feature is also reachable from Claude, Cursor, or Claude Code through the Model Context Protocol (MCP), so your AI agent can read and write the same board you do.
Solo founders, indie hackers, and small teams (up to ~10 people) who use AI as part of their daily build loop. If you find yourself re-explaining your roadmap to ChatGPT every session, Taskachu replaces that friction with persistent state your AI can read directly.
Yes. Taskachu exposes a Model Context Protocol (MCP) server at app.taskachu.com/api/mcp. Generate a workspace-scoped Personal Access Token in Settings → API Tokens, paste the snippet into your Claude Code, Claude Desktop, or Cursor config, and your AI agent can list, create, update, move, and delete cards and documents in that workspace — same surface you use in the UI.
Free tier covers one workspace with a capped monthly AI quota — plenty for an early-stage solo project. Pro is $12 per month, includes higher AI quotas, multiple workspaces, and a 7-day free trial with no credit card required.
Yes. Each workspace has an Imports section that accepts Jira XML exports, ClickUp CSV / JSON, and Confluence HTML / ZIP exports. Imports are two-step: analyse first to see what will be created, then execute when you're happy with the plan.
Task decomposition (turn a card into subtasks), card prioritisation, brainstorm sessions that synthesise into board structures, daily standup planning with conversational follow-up, weekly retrospective generation, agent-export packets for handing a card or board to an external code-gen tool, and Ask Taskachu — a workspace-scoped retrieval-augmented assistant that answers questions about your own cards and uploaded documents.
Yes. Go to Settings → Account → Delete account. We send a confirmation link to your email; clicking it permanently removes your account, every workspace you own (with cascade to all boards, cards, documents, and integrations), and cancels any active Stripe subscriptions. The process is irreversible by design.
