Try for free

A product manager in your terminal that remembers your project

Give tasks straight from Claude Code or OpenCode. The AI Product Manager agent keeps your project context, shows what it relies on and learns from your edits.

Try for freeHow to connect

How it works

Three steps from sign-up to your first task

  1. 1

    Get a key

    Sign up and create an API key in your dashboard.

  2. 2

    Connect a client

    Add the key to Claude Code or OpenCode.

  3. 3

    Give tasks

    The agent answers like a product manager and remembers the project between sessions.

Roles for your tasks

Each role covers its own categories of work

  • Product manager

    Takes a product from idea to release.

    • Roadmap and priorities
    • Requirements and PRDs
    • Product metrics

  • Analyst

    Digs into data and states conclusions.

    • Market research
    • Competitor analysis
    • Hypotheses and A/B tests

  • Backlog owner

    Keeps the work queue in order.

    • Task grooming
    • User stories
    • Acceptance criteria

  • Documentation writer

    Writes clear text for the team.

    • Release notes
    • Specifications
    • Onboarding

You see every step

The agent never works blind: context and decisions stay visible

  1. 1

    Task

    You state the task in your terminal.

  2. 2

    Project memory

    The agent pulls the relevant facts and rules.

  3. 3

    Answer

    You get an answer and see what it is based on.

  4. 4

    Rule for approval

    If you corrected the agent, it proposes a rule to remember. You decide.

Comparison

The agent, an in-house PM and a freelancer

CriterionAI Product Manager agentIn-house PMFreelancer
PriceSubscriptionSalary and taxesHourly rate
24/7 availabilityYesNoNo
Project memoryVisible and editableIn a person's headLost between gigs

What the work looks like

A terminal with the agent's answer and the memory screen in your dashboard

Terminal

Prepare release notes for version 2.4

Using project rules: business tone, no jargon.

Release notes are ready. I added a “What changed” section.

Rule proposed: add “What changed” to every release note. Save it? [y/n]

Project memory

  • Writing tone: business, no jargonActive
  • Release notes ship on ThursdaysActive
  • Add a “What changed” sectionAwaiting approval

Pricing

Start for free. Paid tiers are coming later

  • Free

    — per month

    • Small starter budget
    • Project memory
    • Claude Code and OpenCode support
    Try for free

  • Standard

    Soon

    — per month

    • More budget
    • All roles
    • Project memory
    Contact us

  • Maximum

    Soon

    — per month

    • Maximum budget
    • All roles
    • Project memory
    Contact us

Frequently asked questions

Can't find an answer? Write to us

Where is my data stored?
Data is stored on servers in the European Union. Staff access is restricted and written to an audit log.
Is my data used for training?
No. Your data is not used to train models.
How is this different from a regular chatbot?
The agent remembers the project across sessions, shows which facts it relies on and proposes rules for your approval. A regular chat starts from scratch every time.
Which clients are supported?
Claude Code and OpenCode. See the “How to connect” page for instructions.
What happens when the budget runs out?
The agent replies with a clear message that the plan budget is used up and shows the renewal date. Your data and project memory stay intact.
Which model does the agent run on?
The AI Product Manager agent runs on commercial language models. We do not disclose the specific vendor: for you it is a single agent, and we take care of choosing and updating the models.