Skip to content

Know

It has read everything. It changes nothing without asking.

Fora AI works from the same graph the product runs on — the posts, docs, goals, projects, tasks and meetings you already have. So the answer comes with its receipts, and the actions it wants to take arrive as proposals you approve or throw away.

Start freeSee pricing

“fora launch” — the decision, the post it was made in, and the task it set in motion.

Answers with receipts

The question a new hire asks in week one — “why did we decide that?” — is usually unanswerable, because the answer is in a thread nobody can find. Fora AI answers it by walking the same links the product draws, and shows its work so you can check it.

  • Reads people, goals, projects, tasks, docs, meetings and threads
  • Cites what it used — the post, the doc, the meeting
  • Streams as it thinks, and shows every lookup it makes
  • Respects permissions: it can only read what you can read
3
NotificationsMark all read
Sarah Chen mentioned you in Launch Sync notes12m
Felix Berg completed Press kit photos41m
Marcus Johnson confirmed How we ship is still true2h
Amara Nwosu assigned you Venue contract countersign3h
Priya Nair commented on Beta signup email #35h

It proposes. You approve.

An assistant with write access to the company is a governance problem, not a feature. So Fora AI doesn't have one. It drafts, and a person decides — which also means the audit trail says who approved it, not which model was running that week.

  • Proposed tasks, docs, posts and nudges — never silent writes
  • Every proposal says what it would do before it does it
  • Approve, and the real object is created with a real owner
  • Discard, and nothing happened
Ship the Q4 Launch
On track Due Oct 31
72%
Launch site live with < 1s load
500 beta signups before announce

Rituals, run quietly

The recurring admin nobody enjoys — writing up the meeting, chasing the goal that went quiet, assembling the weekly summary — is the work an assistant should be doing. Each ritual is off until an admin turns it on, and each one produces proposals, not faits accomplis.

  • Meeting actioner — a transcript becomes proposed tasks and a recap
  • Weekly goal rollup — posted to the company feed
  • Stale-goal nudges — a quiet check-in on a goal that's gone silent
  • Admins turn each one on or off; nothing runs by surprise

Everything in it

  • Named conversations, kept and searchable
  • Streamed answers, token by token
  • Visible tool activity for every run
  • Reads people, goals, projects, tasks, docs, meetings and threads
  • Answers cite the objects they came from
  • Proposals: tasks, docs, posts and nudges
  • Human approval before anything is written
  • Standing memories you can read and delete
  • Meeting actioner ritual
  • Weekly goal rollup ritual
  • Stale-goal nudge ritual
  • A DM with Fora for quick asks in the flow of chat

Questions

Is our data used to train a model?
No. Your workspace is used to answer your questions and nothing else.
Can it see documents I haven't shared?
No. It runs with your permissions. If a doc wouldn't appear in your search, it isn't in the answer either.
What stops it doing something wrong?
It can't act on its own. Everything it wants to change arrives as a proposal with a plain description, and a person approves it. The worst case is a bad suggestion you delete.
Does it remember things between conversations?
Only what it's been told to remember, and you can list those memories and delete any of them.

Keep reading

Chat & feedThe company square, group chat, and windows that dockDocs & knowledgeThe handbook and the specs, written down and findableProjects, tasks & goalsBoards, subtasks, and progress that rolls up on its ownMeetings & virtual officeA floor you can walk, knock on, and split into rooms

Run the whole company from one place.

Start free

Free for up to ten people. No card. Bring the company. Leave the stack.