Rindo

Let your AI into the project — not into the controls.

A hosted workspace for software teams — tasks, a markdown knowledge base, test cases and releases in one place. Connect the Claude or ChatGPT you already use: it works under your permissions, has no tools to delete or reconfigure anything, and leaves a record your whole team can read.

Start free

Nothing to operate. Start from one project. No card needed.

Already running AI on your projects? Talk to us

Works with Claude — web, desktop and mobile · ChatGPT · Claude Code

Sign in from your assistant, or use a personal access token. Sign-in access is read-only unless you grant write.

Under your rules

What your AI can't do

There are no delete tools. No settings, workflow or membership tools either. Over MCP those actions are not denied — they are absent. Whatever a prompt tells your assistant, it has nothing to remove your work with, and nothing to rewire your project with.

What your whole team can see

Every write — by a person in the app or by an assistant over MCP — lands in the project's audit trail with who did it and through which channel. Any project member can open it. Entries are never edited.

What waits for a person

Reviews and translations from your AI always arrive as proposals, with a diff. Flip one project switch and every edit it makes to an existing page waits for approval too. Approval is refused if the page changed underneath — a newer human edit is never silently lost.

How it works

  1. 1

    Connect the AI you already use.

    Add https://app.rindo.io/mcp as a connector in Claude or ChatGPT and sign in — or hand Claude Code a personal access token. Sign-in access starts read-only.

  2. 2

    Ask, file, draft, translate.

    Your assistant reads tasks, specs and test cases, files work, drafts pages and proposes changes — under your permissions, and by the workflow rules your project set. It can read those rules before it acts.

  3. 3

    Review what matters.

    Proposed edits wait on the page with a diff: approve or reject. Everything else is on the audit trail.

A knowledge base your AI can edit without breaking it

Every page is plain markdown to your AI, written in a block editor with a slash menu. What you see is what your AI reads — so its edit comes back as a clean diff: one changed paragraph, not a reshuffled page.

  • Markdown in, markdown out. Bring a zipped folder of markdown with its tree intact, previewed before anything is written. Take any page out as Markdown, PDF or Word.
  • Files become pages. Upload Word, Excel, PowerPoint or PDF files; Rindo turns them into pages and tells you exactly what it could not carry over.
  • Diagrams are text. Your AI writes a Mermaid block; your team sees the picture; the diff stays readable.

From spec to release

BA

Paste the meeting notes and ask for a first spec. Then ask for a review: findings come back with a severity, and the rewrite arrives as a proposal you can diff.

Developer

In Claude Code, your AI reads the task, the project guidance and the linked code before it starts. Branches, commits and pull requests that carry the task key link themselves; merging the PR can move the task. (after you connect your GitHub or GitLab repository)

QA

Ask for test cases from a requirement, run them, record the results. A release's readiness counts the tests that passed on the work linked to it.

PM / PO

"How did the last sprint go?" "What is in 2.1?" The answers come from the record — velocity, burndown, changelog — not from anyone's memory.

Translator

Ask for the Japanese version of a spec. It arrives as a proposal beside the source, and the page tells you when it falls out of date.

Give every seat a playbook

Name the roles your project needs — BA, QA, reviewer — each with its own instructions. Whoever holds the seat, a person's assistant or an agent, works from the same page; in Claude the holder picks the role as a prompt.

One page, four languages — and you always know which one is out of date

Ask your assistant to translate the spec into Japanese. The translation arrives as a proposal, side by side with the source; a bilingual reviewer approves it. When the English changes, the Japanese page is flagged Outdated on its own — nobody has to remember.

Rindo doesn't translate. It keeps translations honest: each language has its own history, and readers switch language in the page header — English, Japanese, Vietnamese, Korean. Export follows the language you are reading, fonts included.

Bring the AI you already pay for

Rindo runs no model and sells no AI credits. There is no built-in chatbot to buy: your Claude or ChatGPT is the assistant, on your own subscription. The only AI call Rindo makes is turning text into search vectors.

What's inside

  • Tasks, subtasks and epics
  • A board that refuses moves your workflow doesn't allow
  • A Triage queue
  • Your own statuses and transitions
  • Sprints with velocity and burndown
  • Dependencies
  • Custom fields
  • Components
  • Test cases, runs and coverage
  • Releases with changelog and readiness
  • GitHub and GitLab links
  • Documents with templates, pins and inline comments
  • Revision history, diff and revert
  • AI review rounds
  • Search in four languages
  • Inbox and watching
  • Owner / Admin / Member / Viewer
  • Interface in English, Japanese, Vietnamese and Korean

Questions

Which AI tools work with Rindo?

Claude on web, desktop and mobile, ChatGPT, and Claude Code. Other MCP clients can connect with a personal access token.

Can my AI delete or break things?

It has no delete tools and no settings tools. Pages are versioned, so any edit can be compared and rolled back, and every write is on the audit trail. A personal access token is your own key: scope it — read-only, or only the capabilities you choose — before you hand it to a coding agent.

Does Rindo read my content with AI?

Rindo runs no model. For search it sends the titles and text of tasks, pages and test cases, and your search query, to OpenAI's embeddings API. Comments, chat and attached files are never sent; a file you turn into a page is a page.

Does Rindo replace Jira or Linear?

Not if you need due dates, timelines, time tracking, automations or a native mobile app today. Rindo is for software teams that want their AI working inside the project record — under rules, with a trail.

Is there a mobile app?

No native app. From your phone you can talk to your project through the Claude app.

Can I get my content out?

Any page exports as Markdown, PDF or Word. The project's Owner can export the whole project as one archive.

What does it cost?

There is a free plan with one project. Team is priced per person, billed monthly, and set up with us. Rindo sells no AI credits. See pricing.

Can we run it ourselves?

Self-hosted deployment for enterprise — talk to us.

Ideas become running systems.

Start freeAlready running AI on your projects? Talk to us