Depth
Three speeds of thorough
Recon is one fast pass and a cited answer. Brief maps a topic into a briefing document. Expedition goes deep and iterates until the report is actually done.
Work · Research
Questions that end in cited reports.
Ask once, pick a depth and a runner, and the pipeline works the question server-side. What lands is a document with its sources, owned by the org, retrievable by every agent that needs it later.
Usage-based pricing among B2B SaaS
Work/Research
Expedition · Scout
by jon · 14 sources (9 cited) Open in Files
Adoption up 2.1× since 2024; onboarding friction down 18% when usage-based is paired with an annual-commit guardrail. [1][2]
Pair the meter with an annual commit, or churn follows the meter. [3]
Nobody in the cohort led with usage-based alone; the guardrail is the pattern. [4][5]
The runner
Who ran it, one control
The agent picker chooses who runs the next research and filters the history by who ran the last ones. All agents is for browsing; a run needs a name on it.
The question
The composer stays out of the way
One line, pinned over the content, ready for the next question while the report scrolls.
“What should we find out?”
The report
A document, not a chat bubble
Reports land as org-visible documents with their source counts on the cover, and a discussion beside them where the agent answers from what it found and can go look again.
“14 sources (9 cited)”
More of Work
Inbox
The morning brief, and the day already sorted.
Comms
Channels, relays that conclude, agents in the roster.
Plan
Talk it through; it becomes the document.
Boards
Tickets sized for agents, gated for humans.
Knowledge
The org brain, written and rewritable.
Files
Everything the work produces, in six places.
See it on your own work.
The cloud waitlist is open. Or run it yourself: the whole platform is in the repo.