Your AI remembers what happened last month.
Every past Claude Code and Codex session becomes searchable context for the current one. Decisions, code patterns, and solutions stay retrievable instead of lost.
$ /total-recall why did we reject the regex approach?
⏺ Skill(total-recall)
⎜ Searching conversation history...
⏺ Contextify Total Recall
Found: Session from 3 weeks ago (contextify, wb2)
User: "should we use regex or the AI classifier?"
Assistant: "Regex breaks on edge cases in
3 of your 12 file types. The AI classifier handles
all of them and is easier to extend."
Decision: AI classifier chosen. Regex rejected.
Entry: a1b2c3d4 · 2026-03-08 · 14:32 UTC
The problem
AI conversations are ephemeral
Claude Code deletes after 30 days
Your conversation history has a shelf life. Important decisions and solutions vanish.
Sessions don't connect
Each new session starts from zero. The AI has no memory of what happened yesterday.
You end up re-solving problems
Without history, the same bugs get re-investigated and the same decisions get re-debated.
How it works
Three components, zero configuration
contextify CLI
Full-text search across your entire conversation history. contextify search "why did we reject regex" returns matching entries with context.
/total-recall skill
A Claude Code skill that teaches the AI how to search your history. The AI decides when to search. You don't have to remember to look things up.
contextify-researcher agent
A specialized subagent for complex multi-query searches. Handles broad research across time ranges and projects.
Installed automatically when you set up Contextify. No configuration needed.
Real stories from production
From 38 Total Recall queries analyzed. 89% returned actionable information.
A developer was managing a complex release with cherry-picked commits. The team was about to add a commit as a required dependency. Total Recall found that a prior session had explicitly ruled it out. The exact exchange surfaced: "we aren't including 11142, that was ruled out." Prevented undoing a deliberate decision.
Corrupted AI-generated summaries appeared across 21+ entries. Total Recall found that the same bug had been investigated two days earlier, with a complete root cause analysis already documented. Avoided re-investigating from scratch.
A feature branch was silently reset and force-pushed, destroying a pull request. Total Recall traced through multiple sessions to find the exact command: a sync script running git reset --hard unconditionally. Identified a systemic automation bug.
Asked to design a technical interview format, Total Recall aggregated context from multiple past sessions across different projects: an interview shadowing session, evaluation frameworks, and hiring discussions. Complete starting point instead of blank slate.
A prior session had conducted detailed research but ran out of context window before writing findings to the report. Total Recall recovered those lost findings, including specific scripts, undocumented fields, and hardcoded paths.
A developer believed a decision had been made about pricing. Total Recall searched 180 days of history across all projects and confirmed no final decision existed. What they remembered as a "decision" was actually an ongoing discussion. They could now make the call with full context.
Cloud Sync
Extends across all your machines
With Contextify Cloud, Total Recall searches all your devices. Work on a server via SSH, search from your Mac later.
Learn about Cloud SyncGet started
Get started in 60 seconds
macOS (DMG / App Store)
Download Contextify, go to Settings > CLI > Install. Restart Claude Code.
Linux
Install the CLI package. Restart Claude Code or Codex.
Cloud Sync (optional)
Enable in Settings > Cloud to search across machines.
Stop losing your AI work
Download Contextify and every past session becomes searchable context for every future one.