{"slug": "whatever-recall-local-ai-knowledge-embedded-directly-within-your-code", "title": "Whatever Recall – Local AI-knowledge embedded directly within your code", "summary": "Whatever Recall launches a developer tool that embeds AI-accessible knowledge directly into codebases without modifying source files, using commit-time stamping and local SQLite lookups to provide instant context for AI coding assistants. The self-hosted tool aims to free AI from human-imposed limitations by giving projects memory and traceable decision history.", "body_md": "# Code that\n\nknows *itself*.\n\nNever changes your code · self-hosted · no card needed · zero tokens to recall\n\n## Stop making your AI think like a human.\n\nA human dev is bound by what they can hold in their head — so we built tools for* human* limits. An AI isn't. It can hold the whole graph, re-verify a claim in 0.25 ms, and never tire. What it lacks is a project that **remembers**: the why behind every line, the decisions it must not undo, the real names of things.\n\nrecall gives it exactly that — a codebase with memory, a conscience, and a track record. Free the AI from human restrictions, give it the context only the project can hold, and it codes the way it actually can: creative, exact, and fast. **The AI are the hands; the project is the coworker.**\n\n[One idea, six consequences — follow the chain](#chain-②)\n\n[NewA letter from Christian — why recall exists](/letter)\n\n“…but what does it do to my code?”\n\n## Nothing. recall never changes a single line.\n\nIt only **reads**. Everything recall knows lives in a local `.mind/`\n\nindex and in your commit messages — never in your source. No reformatting, no refactor, no edits, no files moved. Delete `.mind/`\n\nany time and your repo is byte-for-byte what it was.\n\n## The code is the only truth.\n\nKnowledge lives beside the code — in a wiki, a ticket, a chat window. The code itself stays flat and silent, so the instant it changes, the doc starts to lie.\n\nrecall keeps the knowledge **in the code**: pinned to the commit that changed it. One source of truth, not two that drift apart.\n\nOne truth in the code — so everything below follows from it.\n\n[Read: why it all connects](/docs/why-it-connects)\n\n## Captured when the AI still knows.\n\nThe expensive thinking happens while the bug is being fixed — then it evaporates. Later, someone (or some AI) has to re-derive the why from cold syntax.\n\nWhile the AI works and the context is still in its head, recall stamps the decision + the why + the anchors onto the commit. Cheap, exact, one sentence.\n\n``` bash\n$ git commit -m \"switch seat check to an atomic RPC\n\nRecall-Why: the JS check-then-insert raced under load and\ndouble-booked the last seat; the RPC locks the org row.\nRecall-Anchors: seats, confirmSeatOrRollback, orgs.ts\"\n```\n\nThe why is written at the knowing moment — so reading it back needs no model.\n\n[Read: write-time stamping](/docs/how-it-works)\n\n## A reader so dumb it's free.\n\nEvery “why is this here?” makes the AI grep and re-read whole files — tens of thousands of tokens per question, paid again every session.\n\nBecause the meaning was written at write-time, reading is a plain SQLite + FTS5 lookup — no model, no embeddings. The finished answer in sub-milliseconds.\n\n- grep the codebase for the symbols\n- open + read 3 candidate files top to bottom\n- re-derive the reasoning from scratch — every turn\n\n- one SQLite lookup over the stamped anchors\n- returns the decision + the why + the links\n- pinned to a commit — traceable, freshness-checked\n\nOrienting is free — so you can afford to read the whole picture, every time.\n\n[Read: the architecture](/docs/architecture)\n\n## Six answers, before the first keystroke.\n\nAn AI opens a file blind — it can't see the open task wired to it, the decision it's about to undo, or what its change will break three files away.\n\nOne `recall brief`\n\nanswers six questions at once — what · where · why · what breaks · open tasks · the trail — read-only, 0 tokens, the full profile of a file.\n\n[1Whatits identity & role](/docs/six-dimensions)\n\n`kind: api-route · POST`\n\nread more →[2Wherethe file, its symbols, its territory](/docs/stamps-and-edges)\n\n`orgs.ts · confirmSeatOrRollback()`\n\nread more →[3Whythe decisions & the SHA behind them](/docs/why-it-connects)\n\n`“atomic RPC — the JS check raced” @a1b2c3`\n\nread more →[4What breaksthe blast radius — dependents & deps](/docs/architecture)\n\n`7 depend on this · leans on 3`\n\nread more →[5Open tasksthe standing intent wired to it](/docs/six-dimensions)\n\n`▸ TASK: migrate seats to RPC (open)`\n\nread more →[6The trailthe path to understanding the subsystem](/docs/self-healing)\n\n`billing → seats → license gate`\n\nread more →The whole picture is affordable — so an AI never silently undoes a decision it never saw.\n\n[Read: the 6 dimensions](/docs/six-dimensions)\n\n## The code knows what it's called. Your AI doesn't.\n\nAn AI invents a symbol name from its training — `enforceSeats`\n\n? `seatLimit`\n\n? — greps this repo, misses (blind grep lands **0 of 12**), and burns the loop guessing again.\n\nrecall doesn't make the search better — it turns it around. It already wrote down this repo's real vocabulary at write-time, so it corrects the guess into the real name before the grep runs.\n\n- invents a plausible name —\n`getUserSession()`\n\n- greps this repo for it —\n**0 hits** - guesses again, re-reads files, burns tokens\n\n- ranks the guess by the repo's own vocabulary\n- corrects it —\n`getUserSession()`\n\n→`resolveActiveSeat()`\n\n- the grep lands first try, no tokens wasted\n\nThe idea is straight from the support trenches: a senior agent doesn't answer faster, they make sure the question is asked so it can't be misunderstood. recall does that for your AI — it structures the question against the repo's real words before a single token is spent guessing.\n\nWe fix the question, not the answer — the code becomes the senior support agent for the AI.\n\n[Read: search-inversion](/docs/search-inversion)\n\n## The memory gets truer the more you work.\n\nDocs rot because nobody notices when reality diverges from them. Six months in, most of the wiki is wrong and nobody trusts any of it.\n\nEvery note is pinned to a SHA. When the code beneath it moves, recall flags the drift (🟡) and offers the fix at the next commit — you approve, the truth stays current. It only flags, never silently rewrites.\n\n- nothing checks the note against the code\n- the AI reads the rotten version\n- and confidently builds the wrong thing\n\n- SHA-pinned, so drift is detectable\n- recall offers to heal or supersede\n- you approve — the old decision stays in history\n\nDrift is caught and healed — and that feeds back into write-time, so the brain compounds. ↻ the loop closes — every commit makes the next one smarter.\n\n[Read: self-healing & the loop](/docs/self-healing)\n\n## The token bill is\n\nthe *whole* story.\n\nFor an AI, context space is the scarce resource. grep finds *where a word sits* — to learn *why*, it must read whole files back into context, tens of thousands of tokens per question. recall returns the decision in ~56. Measured cold-start on two live production repos (a 240-commit app, a 668-commit CMS), no recall trailers planted — that gap is the product.\n\n*×*\n\n*×*\n\n**Three knowledge questions, one production repo** measured · head_to_head.py\n\n| Method | Time | Tokens | The why? |\n|---|---|---|---|\nrecall() | 2.18 ms | ~152 | ✓ direct |\n| git grep + read files | 147 ms | ~214,000 | ✗ only where |\n| code-index search | ~seconds | — | ✗ 0 hits · 2d stale |\n\n**Tokens per question · grep vs recall** demo\n\n| Question | grep+read | recall |\n|---|---|---|\n| Why must RLS writers set workspace_id? | 48,200 | 56 |\n| Why split the viewer from the editor? | 71,500 | 61 |\n| Why render the search modal via createPortal? | 39,800 | 52 |\n\n**Latency scales flat** measured · FTS5\n\n| Repo | Anchors | Median | Max |\n|---|---|---|---|\n| fixture | 1,204 | 0.04 ms | 0.09 ms |\n| production app | 43,722 | 0.41 ms | 0.67 ms |\n| production CMS | 108,627 | 0.68 ms | 1.0 ms |\n\n## The whole chain,\n\npaid back in your bill.\n\nSix links, one outcome: the AI finally has what it needs to do its best work, and the company stops paying for it to guess. Here's what lands on the balance sheet.\n\n### Millions of tokens saved\n\n152 vs ~214,000 tokens for the same three answers, measured. Across a team, across a year, that read-time saving scales into the millions — every orientation that used to cost a model call is now free.\n\n### Code that gets better on its own\n\nThe loop compounds: every commit stamps new knowledge and heals old drift, so the most-worked code carries the truest memory. The longer it lives, the smarter it gets — passively.\n\n### Faster results, fewer mistakes\n\nThe AI is briefed before it edits and warned before it repeats a known mistake — so it stops re-deriving, stops hallucinating constraints, and stops undoing decisions it never saw.\n\n### Searches that land first try\n\nSearch-inversion fixes the guessed name to the repo's real vocabulary before the grep — no more guess → miss → guess loop burning tokens and wall-clock.\n\n### Sharper analysis & onboarding\n\nBlast radius, contested hotspots, the causal trail — and a new dev or a fresh agent inherits the whole repo's reasoning in one command.\n\n### Local, offline, one tool\n\nYour code and memory never leave your machine — no cloud, no telemetry, no vector DB, no API key. One small tool, pure stdlib, runs with the network cable pulled.\n\n## Your code never\n\n*leaves* your machine.\n\nrecall is a small, self-hosted tool — not a platform. It needs exactly one thing: access to your project folder. That's the whole footprint, and the whole setup takes two minutes.\n\n### Runs on your machine\n\nA small CLI and a local index inside your repo (.mind/). No server to run, no agent in the background. Works offline between hourly licence checks (a brief online confirmation, with a short grace window) — your code never leaves your machine.\n\n### No data sync. Ever.\n\nNothing is uploaded, nothing is synced, no telemetry. It only reads your project — the website just checks your license.\n\n### Open code — maximum trust\n\nThe full source is public on GitHub. Read every line that touches your repo before you run it.\n\n``` bash\n$ pip install git+https://github.com/heidrich/whatever-recall.git\n$ recall init .\n# indexed — your repo is its own memory now\n```\n\nThat's it. No config, no cloud account, no upload — the memory lives in your repo and ships with every clone. Questions? The [source on GitHub](https://github.com/heidrich/whatever-recall) is one click away.\n\n## Your decisions never\n\n*leave* your machine.\n\nThe most sensitive thing about a codebase is the *why* behind it. recall keeps it in a local brain, not in your source — so your code ships clean, and your reasoning stays yours unless you choose to share it.\n\n### The why lives in the brain, not your code\n\nEvery decision you stamp sits in the local .mind/ index — beside your code, never inside it. Your source ships byte-for-byte clean; there's nothing to scrub before you publish.\n\n### Per-note visibility: team or private\n\nMark a note --private and it's yours alone — never in an export, never in a shared brain. Notes can only get more private, never less. You can't widen something by accident.\n\n### Share a brain — without leaking\n\nrecall export writes a shareable copy with every private note stripped out. Whether your team shares it is your call; what's in it is recall's guarantee: team knowledge, never private.\n\n### Two guards, both fail-closed\n\nThe export aborts rather than write a brain that still holds a private note, and check-leak blocks a commit that would stage one. A private decision physically can't reach a teammate's clone or a public repo.\n\n``` bash\n$ recall stamp \"why we chose this\" --private\n  🔒 private — stays in this brain\n\n$ recall export --out .mind/shared.db\n# private notes stripped · gate verified clean\n```\n\nA solo developer ships a clean codebase by construction. A team shares the knowledge it wants to — and nothing it doesn't. Your code is the public artifact; your reasoning is yours. [How private knowledge stays local →](/docs/privacy)\n\n## One product.\n\nPriced by *seats*.\n\nNo tiers, no feature locks, no “upgrade to unlock.” Everyone gets every feature — the CLI, the dashboard, Power Mode, the web-AI bridge, MCP, unlimited repos. You only choose how many people share the memory. Buy a single seat and run solo, or add seats and the team tools appear. Try all of it free for 14 days — no card.\n\n1 seat = run solo. Add a 2nd to unlock team tools (members, keys, billing).\n\n[Start 14-day trial→](/account)\n\nNo card. Every feature. Cancel anytime.\n\n- The CLI, the dashboard, MCP & git hooks\n- Power Mode & the web-AI bridge\n- Token-free recall — reads cost zero model tokens\n- Works offline between hourly seat checks (a short grace if a check can't reach us)\n- Unlimited repositories, unlimited stamps\n- Self-hosted — your code & memory never leave your machine\n- Team tools at 2+ seats: members, license keys, shared billing\n- Every new release becomes open source (Apache 2.0) after 3 years\n\nThe same product whether you're one developer or a hundred — the only thing that scales is the number of people sharing the memory.\n\nI firmly believe that education — and exploring and debating problems together — is the best way to lead humanity into a fantastic future.\n\nSo schools and universities use recall for teaching and academic research at **no charge — no fees, ever**. We'd love to hear what you're building. The one honest line: if the research becomes a product or spin-off that earns money, it needs a regular plan from that point on.\n\n[Get in touch](mailto:support@mccain-digital.com?subject=Education%20%2F%20research%20use%20of%20recall)\n\n14-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. Licensed under the Business Source License 1.1 — **every released version becomes open source (Apache 2.0) three years after its release**, so your stack can never end up in a proprietary dead end.\n\n“We document constantly, then change the plan and never update the doc. The memory rots, the AI reads the rotten version and builds the wrong thing. So we moved the truth into the code itself — the one thing that can't lie.”\n\n[McCain Digital](https://mccain-digital.com)· Bavaria ·\n\n[read the full letter →](/letter)\n\n## A smart AI greps a name it invented.\n\nSmart, self-aware code knows it.\n\nrecall fixes the guess to your repo's real vocabulary before the search — because it's the only thing that recorded what you actually call things.\n\n## It's time to unlock the full power of AI coding.\n\n**The foundation is real today** — the code becomes self-aware, reading is free, the search stops guessing. From there we harden on your feedback, then open it up into a real app with deeper features. **The horizon we're building toward** (not shipped yet, and we'll always say so plainly): today your code lives in folders and files built for *human* reading. An AI navigates the graph, not the tree — and recall already holds that graph. The idea turns it around: the graph as the truth, the file layout as a view rendered from it, so an AI can structure code the way it thinks while recall renders it back to ordinary files for your compiler and git. It starts small — recall *suggesting* where code really belongs.\n\n[Read where recall is going](/docs/evolution)", "url": "https://wpnews.pro/news/whatever-recall-local-ai-knowledge-embedded-directly-within-your-code", "canonical_source": "https://whatever-recall.com", "published_at": "2026-06-19 11:05:44+00:00", "updated_at": "2026-06-19 11:37:38.214908+00:00", "lang": "en", "topics": ["developer-tools", "ai-tools", "ai-agents", "ai-infrastructure", "machine-learning"], "entities": ["Whatever Recall", "SQLite", "FTS5"], "alternates": {"html": "https://wpnews.pro/news/whatever-recall-local-ai-knowledge-embedded-directly-within-your-code", "markdown": "https://wpnews.pro/news/whatever-recall-local-ai-knowledge-embedded-directly-within-your-code.md", "text": "https://wpnews.pro/news/whatever-recall-local-ai-knowledge-embedded-directly-within-your-code.txt", "jsonld": "https://wpnews.pro/news/whatever-recall-local-ai-knowledge-embedded-directly-within-your-code.jsonld"}}