A wiki for AI-assisted teams

Institutional memory for AI-assisted development.

Track the decisions, citations, and rationale behind every code change so your team's knowledge doesn't get lost in chat logs and AI session transcripts.

Get started Sign in →

Screenshot of Hintertide — the SPA's page view with the document tree, page content, and proposals queue

What it is

Hintertide is a wiki built for codebases where AI is doing real work. Every claim links back to a specific file at line precision; every page tracks its own confidence and freshness.

AI tools — Claude Code, Cursor, ChatGPT MCP, anything that speaks MCP — propose changes against the wiki, citing the source. Humans review and accept. The institutional memory accumulates instead of being lost in a thousand AI session transcripts.

Citations at line precision

Every claim points to specific lines in specific files. No "according to the codebase" hand-waving.

Confidence + freshness

Each page tracks how verified it is and when its sources were last checked. Stale claims surface visibly.

AI proposals + human review

AI suggests; humans approve. Accepted changes carry full attribution forward — who proposed, who reviewed.

Search and reference

Look things up from inside an AI session via MCP, or browse the wiki at app.hintertide.com.

Why you need it

AI tools generate code at unprecedented speed. The decisions behind that code scatter across chat logs, terminal sessions, and Slack threads — and disappear.

Sessions are ephemeral

An AI agent figures out why your auth middleware is shaped a certain way. Five minutes later, the next session re-derives it. The understanding never accrues.

Conversations are unsearchable

"We decided to skip GraphQL because of N+1 issues, remember?" "Which channel?" Decisions don't travel between AI tools, between team members, or across time.

Knowledge gets re-discovered, badly

Without durable memory, your team's hard-won understanding has to be re-extracted from the codebase every time someone — or some AI — needs to act on it.