write it down.the shared workspace for teams working with AI agents
Tatara holds what your company knows in one place your people and your agents both read and write. Plain Markdown underneath, open standards throughout, so nobody has to re-explain the company again.
- ❯ we’ve dropped the 50-person ceiling. update the ICP
- ⏺ tatara – search_documents (MCP)query: "ICP"
- ⎿ 1 result · Sales/ICP.md · Ideal customer profile
- ⏺ tatara – get_document (MCP)path: "Sales/ICP.md"
- ⎿ reference · sales, sales/positioning, acquisition · 1 link
- ⏺ Changing “companies of 10–50 people” to “companies of 10 people and up” under Who this is for.
- ⏺ tatara – write_document (MCP)target: Sales/ICP.md · body
- ⎿ status: committed
ask any agent,
get your company’s answer
Claude, ChatGPT, Gemini and Claude Code read the same docs your team wrote, so the answer comes from what you actually decided, not from a guess.
- ❯ does a two-person agency fit our ICP?
- ⏺ tatara – search_documents (MCP)query: "ICP"
- ⎿ 1 result · Sales/ICP.md · Ideal customer profile
- ⏺ tatara – get_document (MCP)path: "Sales/ICP.md"
- ⎿ reference · sales, sales/positioning, acquisition · 1 link
- ⏺ Not yet. The ICP disqualifies teams with fewer than three people who write anything down, and it needs an agent already in production. Source: Sales/ICP.md.
connect any agent in a minute
Add Tatara once in each tool you already use. Every agent reads the same current docs, via MCP.
Claude
ChatGPT
Gemini
Paperclip
OpenClaw
Hermes
one doc, no export
The page your team edits is the page the agent reads. No sync, no second copy, nothing to keep in step.
your agents can save to it too,
and they have to do it right
Ask an agent to capture a decision or update a playbook and it lands in the right folder with a title, a summary and tags. No stray “Notes” pages, ever.
- ❯ save a note: discovery calls are 30 minutes from now on, not 45
- ⏺ tatara – get_taxonomy (MCP)
- ⎿ folders: Sales/, Sales/Decisions/ · tags: sales, sales/discovery, process …
- ⏺ tatara – write_document (MCP)Sales/Decisions/discovery-call-length.md · type: decision · tags: sales, sales/discovery
- ⎿ invalid_input · description is required — one plain sentence stating what the document is.
- ⏺ tatara – write_document (MCP)+ description: "Discovery calls are 30 minutes from 2026-09-23."
- ⎿ status: committed
nothing half-filed gets in
A save missing its title, summary or tags is refused with a plain error. The agent fixes it and tries again.
your folders, your tags
Agents read your folder tree and tag list before they save, so what they file lands where your team would have put it.
a doc when you read,
a board when you plan
Turn any document into a filterable base, a table with formulas, a kanban board or a task list. Same file underneath, nothing converted.
| Title | Type | Tags | Updated |
|---|---|---|---|
| Ideal customer profile | reference | sales, acquisition | 2026-07-21 |
| Pricing playbook | playbook | sales, pricing | 2026-07-18 |
| Discovery call script | process | sales, discovery | 2026-07-11 |
plain Markdown underneath
Every document is a file you could open in any editor. The links between them form a graph you can see.
built for a team
Roles, invitations and unlimited viewers on every plan. The whole company can use it while a few people keep it true.
your knowledge, in the open
Plain Markdown files, one per thing you know, with links between them. No proprietary format and nothing you couldn’t open in any editor, so people stay with Tatara because it’s good, not because leaving is expensive.
free to start. viewers are always free.
Start with 500 documents and no card. Viewers are free and unlimited on every plan, and agents connect on every plan, the free one included.
Starter
$49/mo
- 5 author seats · unlimited viewers
- 10,000 documents
- Unlimited agent access
Team
$149/mo
- 10 author seats · unlimited viewers
- 25,000 documents
- Unlimited agent access
Solo and Enterprise, annual billing, and the full comparison live on the all plans page.
one set of docs
for humans and agents
Write it down once, keep it current together, and stop re-explaining your company to every AI.
I’m Angus. I built Tatara because I was pasting the same company context into every AI chat, and watching good agents produce confidently wrong work whenever I forgot.
It’s early, and I’d rather show you what has shipped than sell you what hasn’t. The changelog says what landed and when. If you’re setting this up for your team, email me and I’ll help.
