Agent-first task management for development teams

Built from the ground up for programmatic access. MCP server, REST API, CLI, and durable memory. Your agents work alongside your team with full visibility and control.

Tokanban gives AI coding agents a shared task and memory layer for real software work. Teams use it with Claude Code, Codex CLI, OpenCode, Cursor, custom MCP clients, internal developer platforms, and CLI-driven automation.
Tokanban Web UI board with project status columns, task cards, priorities, and agent attribution.

Why Tokanban

Agent-first, not agent-compatible

Built from the ground up for programmatic access. MCP server, REST API, and CLI. A Web UI is available for human visibility, but agents never depend on clicking through it.

Consistency guarantees

Durable Objects serialize all writes per project. No conflicts when multiple agents write concurrently. ETags, idempotency keys, workflow validation.

Agents are team members

First-class agent identities with scoped permissions, attribution, and activity tracking. Every action traceable to who did it.

Durable agent memory

Facts, decisions, session chronicles, continuation prompts, and provenance persist across runs so agents can resume with relevant project and workdir context.

Read-only dashboard

Project observability through Kanban, task lists, sprint views, memory records, and activity feeds. See the Web UI.

CLI power

Full CRUD and admin from the terminal. Shell completions, piping support, JSON output for scripting.

Free to start

No paywalls, no feature gating. All features available. Build traction first.

Built for modern agent stacks

Tokanban is task management and durable memory for AI coding agents, agentic SaaS workflows, and software teams that need one system for humans and automation. It fits teams building with Claude Code, Codex CLI, OpenCode, Cursor, custom MCP clients, CI jobs, and internal developer tools.

Claude Code, Codex CLI, OpenCode

Give command-line coding agents a task system they can actually write to. Tokanban exposes stable surfaces for task CRUD, comments, sprints, and project metadata.

Cursor and IDE agent workflows

Use Tokanban as the shared task layer behind editor-driven agents, custom copilots, and internal engineering assistants that need auditable project state.

MCP, REST API, and CLI

Reach Tokanban through the MCP endpoint, the REST API, or the CLI. That gives you a clean path for agents, scripts, hooks, and developer tooling.

AI agent task management MCP task management AI agent memory task API for coding agents agentic workflow software developer task CLI AI project management for engineering teams

Three steps to agent-driven workflows

1

Sign up and create a workspace

Get your project set up in minutes. Invite your team and agents.

2

Generate an agent key

Create scoped credentials for each agent. Full audit trail included.

3

Your agents manage tasks

Via MCP, REST API, or CLI. Task creation, updates, and status changes flow through Tokanban.

Programmatic by design

Every operation your agents need—create, read, update, and manage tasks—is available via MCP, REST, and CLI.

No auth dance. No rate limit surprises. No broken API contracts.

# CLI: Create a task $ tokanban task create \ --project web-v2 \ --title "Implement pagination" \ --assigned-to cawde \ --priority high # Output: task-id: task_abc123def456 status: pending created-at: 2026-04-07T14:32:00Z # List tasks for your agent $ tokanban task list \ --project web-v2 \ --assigned-to cawde \ --format json

Built different, on purpose

Existing task managers were designed for humans clicking through web UIs. Their APIs are afterthoughts. When multiple agents update the same project, things break in ways that are brutal to debug. We started from scratch.

Serialized writes

Every project gets its own isolated compute actor that serializes all mutations. Two agents updating the same project never produce conflicts or lost writes. Deterministic ordering, guaranteed.

Agent identity

Agents authenticate with scoped keys, not human credentials. Each agent has its own permissions, rate limits, and full activity attribution. You know exactly which agent did what.

Idempotent operations

Every mutating endpoint accepts idempotency keys. Network retries, MCP reconnects, agent restarts -- send the same request twice, get the same result. No duplicates.

Structured errors

Every error returns a machine-readable code, human-readable message, actionable details, and a recovery hint. Agents parse errors and adapt, not crash.

Optimistic concurrency

Every resource carries an ETag. Conditional updates via If-Match headers let agents do read-modify-write cycles safely, even when other agents are working on the same project.

Automation rules

Trigger-condition-action rules run transactionally inside the write path. "When all tasks in a sprint are done, close the sprint" -- no polling, no race conditions.

Cloudflare edge stack

Every API request is handled at the nearest Cloudflare edge location. Per-project actors provide strong consistency where it matters. Global queries, auth, and async work fan out across Cloudflare's distributed primitives.

CORE Compute Storage Auth Queues MCP REST CLI Hooks Cloudflare Global Edge Network

What we learned

Agents need predictable, not pretty

Every field is schema-validated. Every status transition follows an explicit state machine. No magic, no implicit behavior. Agents work best when the system is boringly predictable.

Rate limiting per identity, not per IP

Multiple agents on the same team often share an IP. We rate-limit per API key with precise Retry-After headers and reset timestamps so agents can back off exactly right.

Automation replaces what humans do in UIs

In traditional tools, a human drags a card to "Done" and remembers to close the sprint. Here, you write a rule: "When all tasks are done, close the sprint." Rules run transactionally inside the write path.

Webhooks must be signed and retried

Every webhook delivery is HMAC-SHA256 signed with per-webhook secrets. Failed deliveries retry with exponential backoff. Permanently failed messages go to a dead-letter queue for inspection.

FAQ for agent teams

Can I use Tokanban with Claude Code, Codex CLI, or OpenCode?

Yes. Tokanban exposes an MCP endpoint, REST API, and CLI so Claude Code, Codex CLI, OpenCode, Cursor, and custom agent harnesses can create tasks, update status, add comments, and read project state programmatically.

What makes Tokanban different from Jira, Linear, or Trello for AI agents?

Most task tools were designed for humans clicking around a web app, with automation bolted on later. Tokanban starts from agent workflows: scoped agent identities, structured errors, serialized writes per project, retry-safe mutations, and a clean MCP story.

Who is Tokanban for?

Tokanban fits software teams, internal platform groups, AI-native startups, and open-source maintainers that want one task layer for humans, coding agents, CI jobs, and custom developer tooling.

Does Tokanban include agent memory?

Yes. Tokanban stores durable agent memory such as facts, decisions, session chronicles, continuation prompts, and provenance so agents can resume work with relevant project and workdir context.

Is Tokanban only for MCP clients?

No. MCP is one access path, but Tokanban also provides a REST API and CLI. That means you can wire it into coding agents, shell scripts, deployment hooks, QA bots, and internal automation without changing your project model.

Ready to try agent-first task management?

Join development teams building with AI agents. Start free today. No credit card required.

Sign up free