Indrasol LogoIndrasol

    Where the work behind your goals gets done.

    Plan projects, goals and milestones, and track every task to done, with your people and your AI agents working in the same place, and a person always in charge.

    Indrasol runs its own work on TasKadence.

    Interactive exampleWebsite relaunch
    Goal: Launch the new site
    33% · 1/3 tasks
    T241328Accessibility auditR
    T856830Build pricing pageR
    T768427Write launch announcementR
    Unlimited
    Projects, tasks, members and teams
    Every view
    List, board, goals, milestones, timeline and sprints
    9 + any
    AI tools connect over one MCP server
    3 ways in
    REST API, Python SDK and MCP
    How it fits together

    From intent to done.

    People plan in the web app. AI agents work through the API, SDK and MCP. Both meet at the tasks, and progress rolls up from there.

    1. 01

      What you want to achieve

      The outcome you are after, written down as a project and its goals.

    2. 02

      Projects

      Where a body of work lives: its people, goals, milestones, sections and tasks.

    3. 03

      Goals & milestones

      Goals say what success looks like; milestones mark the dates on the way.

    4. 04

      Tasks

      The work itself, done by people and AI agents alike.

      PeopleAI agents
    5. 05

      Progress

      Rolls up as tasks move: goals, milestones, sprints and the Timeline.

    6. 06

      Done

      Tasks complete, the milestone closes, the goal is met.

    Peoplethe web app: lists, boards and timelines in the browser
    AI agentsthe REST API, the Python SDK or the MCP server
    Plan and see progress

    Every view of the work, from the same tasks.

    Name what success looks like

    Each project's goals on one page, with the tasks under them and percent complete that rolls up as work is done. Switch to a timeline to see every dated goal at once.

    • Progress rolls up automatically
    • List or timeline view
    • Drag a goal to another project, tasks and all
    TasKadence Goals page: projects expanded to their goals, each with its tasks and percent complete

    Also in every workspace: sprints and team capacity, task dependencies, comments with @mentions, attachments, sections, pinned projects and keyboard shortcuts (⌥N creates a task from anywhere).

    People and agents, safely

    Let AI agents do real work, with a person in charge.

    Agents work where your team works

    Connect an AI tool over MCP, or call the REST API or the Python SDK. Agents create and update tasks, projects, goals and milestones in the same workspace.

    Marked, and held for review

    Every task an agent creates carries a Bot mark. Hold agent tasks for a person to accept or reject: for private projects, for every agent task, or never.

    On the record

    Every agent action is in the audit log, under the token's or the app's name. An agent can never approve its own task.

    Scoped, and revocable

    Read-only mode, tool groups and per-scope consent limit what a connection can do. Revoke a token or disconnect an app at any time.

    Bring the assistant you already use
    ClaudeClaude CodeCursorVS CodeGitHub CopilotCodexGemini CLIWindsurfZedAny MCP client
    TasKadence consent screen: Claude asks to access a TasKadence workspace, with one checkbox per scope, the organization it will act in, and Allow or Deny
    Connecting a tool: untick any scope before you allow it.
    Developers

    Build on the same API the agents use.

    Script it, automate it or plug it into your own tools, with the same API your agents use.

    REST API

    /v1

    Every object your team works with, over JSON: tasks, projects, goals, milestones, teams, sprints and webhooks.

    Python SDK + CLI

    taskadence · tm

    A typed client with pagination, retries and typed errors, and a command line for scripts.

    MCP server

    remote · local

    One server for AI tools: connect over OAuth or with a token, read-only if you like, with the tool groups you pick.

    quickstart.pyPython SDK
    from taskadence import Taskadence

tm = Taskadence()  # reads TASKADENCE_TOKEN

# every task, page after page
for task in tm.tasks.list(org_id="O0001"):
    print(task.task_id, task.status, task.title)
    Built for trust

    Access you can see, scope and take back.

    Security at TasKadence

    Org-bound access

    A token belongs to one organization; an app connection acts in the one organization chosen at consent.

    Scopes and token policy

    Namespaced scopes on every token. Owners can require expiry, cap token lifetime and turn off personal tokens.

    Audit log

    Who did what and when, for people, tokens and apps, with retention you set and CSV export.

    Signed webhooks

    HMAC-SHA256 in the Standard Webhooks format, with retries, a delivery log and replay.

    Data per organization

    Isolation enforced in the API, backed by Postgres row-level security.

    Sign-in

    Email codes with two-step sign-in, Google and Microsoft sign-in, and SAML single sign-on.

    Start free with your whole team.

    Every feature, for every person in your organization. See how storage works on the pricing page.

    FAQ

    Questions, answered.

    More in the TasKadence guide, or talk to our team.

    What is TasKadence?

    A work management app by Indrasol. Plan projects, goals and milestones, and track every task to done, with your people and your AI agents working in the same place.

    How do AI agents work in TasKadence?

    Connect an AI tool over the MCP server, or use the REST API or Python SDK. Agents can then create and update tasks, comments, projects, goals, milestones, teams and sprints, within the access you give them.

    Can an agent's work go live without a person checking it?

    Your organization decides. Agent tasks can be held as Needs review for private projects, for every agent task, or never. Rejecting a task archives it; nothing is deleted. An agent can never approve its own task.

    Which AI tools can connect?

    Claude, Claude Code, Cursor, VS Code, GitHub Copilot, Codex, Gemini CLI, Windsurf, Zed and any other client that speaks MCP, all through one MCP server.

    Will I know when work is assigned to me?

    Yes. TasKadence emails you when a task is assigned to you.

    How do people sign in?

    With a one-time email code and optional two-step sign-in, with Google or Microsoft, or through SAML single sign-on that you set up.

    How do we get started?

    Create a free account, make an organization and a project, add a few tasks and invite your team. The TasKadence guide walks through every feature.

    Put your goals and your agents to work.

    Start free with your team, or book a demo and we'll show you TasKadence with your own projects.

    Get started freeBook a demo