Repos, trackers, wikis and dashboards, indexed where they already live.
Every system you run.
One place to ask.
Connect the repos, trackers and runbooks you already have. Every page lands in one index, and every answer arrives with the source attached.
# CONNECT A SOURCE
## TASK
READ THE REPO BELOW, LIST EVERY SERVICE IT DEFINES, WRITE ONE PAGE FOR EACH, AND LINK EACH PAGE TO THE RUNBOOK THAT NAMES IT.
Paste it into whichever tool already has your repo open.
Connect a source
A source is any system that already holds the truth: a repo, an issue tracker, a wiki, a dashboard. Connecting one puts its pages on the graph.
Nothing is copied. Pages are indexed where they live, so a page that changes at the source changes here within the hour.
A source connected this morning is searchable by lunch.Once two sources are on the graph, it starts drawing the wires between them.
Your docs stop being true quietly
Nothing announces it. A service gets renamed, the runbook keeps the old name, and the page both of them point at is now wrong.
of indexed pages describe a system that has since changed
A hit is not an answer
Both routes are the same length. One ends on a page that still matches its source. The other ends on a guess.
- Arrives cited
- The page still matches the source it was written from.
- Stops short
- A result with nothing behind it, so nobody acts on it.
One graph, wired to keep itself true
When a source moves, the graph proposes the edit for review.
Every page is re-checked against its source on each commit.
People search it. Agents query it. Same pages, same answers.
We stopped answering the same question in chat. It is written down once, and the answer arrives with the page attached.
One graph, and one way in
Every read passes the same gate. Your directory decides who sees a page, and the log says who saw it.
Book a demoSingle sign-on
Your directory decides who reads a page, down to the section.
Group sync
Groups arrive over SCIM, so a leaver loses access the same day.
Audit trail
Every read, edit and answer is logged against the page it touched.
Data residency
Choose the region your index is built in and never leaves.
A stale page used to cost us an afternoon. It gets flagged the hour its source moves, and the draft is waiting.
Four repos and a wiki, indexed in a week.
Support replies now cite the page they came from.
One handbook for two thousand people, kept current daily.
Every service page rebuilt from the code that ships.
Nine sources, one index, and a search that lands.
Runbooks that change when the alert they answer does.
Pay by space and seat
Free
One space, three editors, and the whole index searchable.
- One space
- Three editor seats
- Community support
Team
Per editor. Unlimited spaces, and a check on every merge.
- Unlimited spaces
- Checks on every merge
- Cited answers
- Support in one day
Scale
Your directory, your region, and the log to prove both.
- Single sign-on and SCIM
- Regional index
- Audit log export
What counts as a source?
Anything with an API behind it: a repo, an issue tracker, a wiki, a dashboard, a table in the warehouse.
Does anything get copied?
No. Pages are indexed where they live, and every answer links back to the system that holds the page.
How does an editor seat work?
A seat is a person who can edit. Readers are unlimited on every plan, including the free one.
What happens when a source changes?
The page is flagged within the hour and the graph drafts the edit, which a person still has to approve.
Can an agent query the index?
Yes. The same index answers a search box and an API call, and both return the page they cited.
How do we get our pages out?
Export any space as Markdown with its history and its links, at any time, without asking us first.
Connect one source.
Ask it anything.
The free plan is one space and three editors. It takes about ten minutes and no card.
unsloppable