alpha · claude code

you've built more
than you remember.

Every Claude Code session, compiled into one local SQLite file. Run flex init, then ask anything.

4,500+ sessions · 275K+ chunks · 1 SQLite file
the query

five calls.
one answer.

Claude reads the schema, then assembles the answer in a few targeted queries.

01

reads the schema

learns what's in your database before writing a query.

02

finds the file

across every path it's lived at, even through renames.

03

pulls session history

every session that touched it, in order.

04

searches for context

why it was built, what it replaced, what influenced it.

05

surfaces the conversation

the exact prompt that spawned the file.

Claude reads the schema before it writes a single query.

capabilities

your codebase has photographic memory.

A background worker keeps it current within seconds.

File Lineage

trace any file through every session that touched it: who made it, why, what it became.

Decision Archaeology

find the session where a decision happened: what was considered, what failed, why you landed here.

Knowledge Graph

importance + relationships precomputed, so every query is weighted by what matters.

your projects

every repo. every conversation. one interface.

Grouped automatically. Search one repo or all of them at once.

core-api
533 sessions 15K chunks indexed
agent-runtime
158 sessions 15K chunks indexed
search-engine
90 sessions 4.2K chunks indexed
+ yours
run flex init
local-first

own your data.

One SQLite file on your machine. Nothing leaves it.

~/.flex
$ ls ~/.flex/cells/
claude_code.db
$ sqlite3 claude_code.db "SELECT COUNT(*) FROM sessions"
4547
no cloud no API key no vendor no rate limits

everything you know.
one query away.

one command. no API keys.