Frequently asked questions
The short answers. For setup details, see the documentation; for anything else, contact us.
What is Rationale?
Rationale is organizational memory for engineering teams. It reads the artifacts your team already produces - design docs, tickets, wiki pages, pull request discussions - and maintains a Decision Map: every significant decision, connected to its evidence, tradeoffs, rejected alternatives, and whatever superseded it. When someone asks why something was decided, the answer already exists, with citations.
Who is Rationale for?
Engineering organizations first - teams where decisions live in RFCs, tickets, and pull requests, and where turnover or growth keeps erasing the reasons behind them. The underlying model is domain-agnostic, but engineering is where we start.
What tools does Rationale connect to?
Google Docs, Jira, Confluence, Notion, GitHub, and Dropbox today, with Slack and Linear on the way. Connectors are read-only: Rationale never writes to, edits, or deletes anything in your tools. You can also upload files directly (.txt, .md, .pdf, .docx, and .vtt meeting transcripts).
Is my content used to train AI models?
No. Your content is never used to train AI models - not by Rationale, and not by the AI providers it calls on your behalf. The Decision Map built from your artifacts is your organization's asset.
Which AI model writes the answers?
Your choice, per workspace. Rationale offers models from Anthropic, OpenAI, and Google - the default is Auto, which names exactly which model it resolves to. You can also bring your own provider API keys so that provider's traffic bills to your account and runs under your own agreement, with usage caps and a visible pause if the key fails.
How is Rationale different from enterprise search or a wiki?
Search and wikis retrieve documents; Rationale maintains reasoning. It reads each artifact once, deeply, and keeps the decisions inside it connected and current - so the answer to 'why did we decide this?' is a maintained object with evidence and history, not a pile of search results to synthesize yourself. When a decision is superseded, the old reasoning is preserved, not overwritten.
How does Rationale handle sensitive content?
Every document is screened locally before anything is embedded or sent to an AI provider. Deterministic detectors look for things like card numbers, government IDs, legal privilege markers, and compensation data; flagged documents are held for your review and are never sent to an AI provider unless you approve them. You can also set per-connector deny rules so matching documents are never stored at all.
Can answers show up inside Jira or Confluence?
Yes - a Jira and Confluence client is in early access: a side panel that shows the decisions, evidence, and timeline behind the ticket or page you're reading, and answers follow-up questions in place. There's also a public API, so the same reasoning is one authenticated call away for anything you build. Contact us for early access.
How much does Rationale cost?
Rationale is free during early access - no credit card required. Pricing will be announced before early access ends.
How do I get started?
Sign up, connect a tool or upload a document with real reasoning in it - a design doc, an RFC, a postmortem - and Rationale builds your first Decision Map from it within minutes. The getting-started guide walks through the first session.