14-day trial · every feature · no card
Same features for everyone.
You pick the seats.
There are no feature locks and no gated tiers — every plan is the full product. The only question is how many people share the memory. Start with 14 days of everything, free, no credit card.
01 — Pricing
One product.
Priced by seats.
Every plan has every feature — the CLI, the dashboard, Power Mode, the web-AI bridge, MCP, unlimited repos. You only choose how many people share the memory. Try all of it free for 14 days — no card, no feature locks.
01 · Solo
$10/mo · $100/yr
Every feature. One developer, unlimited repos.
- 1 seat
- CLI · dashboard · MCP · git hooks
- Power Mode & web-AI bridge
- Token-free recall, offline-capable
- Annual = 2 months free
02 · Team
$100/mo · $800/yr
Every feature. One shared memory for the whole team.
- Up to 10 seats — flat, no per-seat counting
- Everything in Solo, same features
- The graph compounds with every teammate
- $8.30 per seat per month at full size
- Annual = 4 months free
03 · Studio / Agency
$250/mo · $2,000/yr
Every feature. For studios running many client repos.
- Up to 25 seats — flat
- Everything in Team, same features
- Unlimited client repos
- Priority support
- 25+ seats? Talk to us
14-day trial, full features, no credit card. Self-hosted either way — your code and your memory never leave your machine; the account only carries your license.
02 — Questions
The honest
answers.
recall is a memory, not a magic box. Here's exactly what it does, where your code goes, and where the limits are.
Does this upload my code anywhere?
No. recall is self-hosted: a small CLI plus a local index inside your repo. There is no data sync, no cloud storage, no telemetry — it just reads your project, on your machine. The website only checks your license. And the source is public on GitHub, so you can read every line that touches your repo.
What happens when the 14-day trial ends?
Your data is never hostage. The .mind folder is yours — reading your memory (recall, briefings, the dashboard) keeps working on the existing index. Writing new knowledge (stamping, healing, Power Mode) needs an active plan. No card is taken for the trial, so nothing renews silently.
Why is recall free of tokens?
Because the thinking already happened at write-time. The reader is plain SQLite full-text search — no embedding model, no LLM in the read path. Spending intelligence once, at the moment of knowing, means reading it back is free and instant forever.
Where does the knowledge live?
In a .mind/ folder inside your project — the index, the plans, the rules. It's deploy-safe (ignored from your build, kept in git) and reproducible: delete it and rebuild from the code. You can't corrupt the project, because the code is the only truth.
Do I have to spend tokens to start?
Never. Bootstrap indexes an existing repo token-free and gives you a working memory in under a second. Power Mode — where the AI reads and stamps richer nodes — is a conscious, reversible upgrade you opt into when you want it.
How is this different from jcodemunch or grep?
grep finds where a word appears; you still read whole files to learn why. jcodemunch maps code structure but its index goes stale silently (we measured it 2 days behind). recall answers the why, pinned to a SHA, freshness-checked — and 1,400× cheaper in tokens.
What if the plan changes after it's recorded?
That's the whole point. recall flags the drift the moment your change contradicts the record, and offers to heal it. You approve, it re-stamps. A sprint plan that's always current — the opposite of an Obsidian vault rotting in the corner.