Shipping in the open

We’re building Lore by running it on Lore.

One engine, two products, built by two people in the open. The engine that powers Lore Pilot and Lore Fleet runs on our own emails, docs, and Slack first, so every entry below shipped against real data, not a demo. Newest work at the top.

Join the Pilot waitlistRequest Fleet access

27 ships · last update July 8, 2026

Tenant-isolatedPostgres RLS, every read

What we’ve shipped

newest first
  1. Engine

    The brain remembers your files, connected

    A task re-fetched a PDF from Gmail in a dozen calls when the file was already sitting in Downloads. Never again: files are now first-class nodes in the knowledge graph, linked to the people, apps, and subjects around them. Ask for "the renters policy pdf" or "the thing Jennifer sent me" and the graph surfaces the file, its path, and the cluster around it. Fast lookups that would have come back empty now escalate to the full graph automatically, a slower right answer instead of a fast nothing.

  2. Engine

    The engine learns how you work, not just who you are

    A third extraction stance, usage mode, turns completed-task debriefs into durable workflow observations: which app you use for which person, how tersely you write to your advisor, that you archive instead of deleting. Your corrections to the agent score highest of all. One-off details are dropped at the same salience gate that keeps the bio clean. These observations are the raw material personal skills compile from.

  3. Engine

    Skills heal themselves when the app changes

    UIs move; playbooks rot. When an agent hits a renamed button mid-task it now recovers from what is actually on screen and files a drift report. A verified fix patches the shared skill, bumps its version, and appends a maintenance log tracing the change to its evidence, the same citation discipline facts have. The runtime overlay heals you instantly; the maintainer makes the fix durable.

  4. Engine

    "What did I ship last year?" finally means last year

    The temporal parser previously did not understand "last year" at all, and "in March" meant "near mid-March". Time references now resolve to exact date ranges, last year, Q2 2025, past 6 months, and facts inside the window always outrank facts outside it. Runs on the dual timestamps every fact already carries.

  5. Onboarding

    Onboarding builds a bio, not a transcript

    Our first sweep faithfully extracted a $4 parking receipt: the authorization code, the eCheck reference, the amount. All true, all garbage for a "who is this person" profile. The engine gated on confidence but not salience. Profile mode fixes it: onboarding now extracts only durable, defining facts, identity, people, routines, work, and drops the transactional noise at a salience floor, with every quality gate (dedup, provenance, citations) still on.

  6. Brand

    One engine, two products

    The site now tells the real story: Lore is the context and trust engine; Lore Pilot runs your day on your Mac; Lore Fleet runs a governed agent workforce over your company's knowledge. Each product lives on its own subdomain, the glass brand (galaxy, aviator, fleet) landed, and every button on the site is real refractive liquid glass, the background bends through it like a water droplet.

    Read the announcement
  7. Backend

    Skills catch when reality drifts from the doc

    Your deployed SKILL.md can quietly stop matching how your team actually works, and agents act on the stale version with full confidence. Lore now regenerates each skill from your current real work and, when the deployed process no longer matches, files a cited report: what the doc says, what your recent Slack threads, emails, and commits show instead, and a one-click patch. Every drift claim links to the exact source it came from.

  8. MCP

    Agents find the right skill by asking

    New MCP tool: lore_skills_search. Instead of dumping the whole skill catalog into an agent's context, the agent describes the task in plain language ("how do we pick a compliance vendor") and gets back just the few most relevant skills, ranked. It loads only the winner. The brain can hold thousands of skills and the agent still pulls a tiny, exact context. Fresh skills get embedded the moment they're emitted, so they're searchable instantly.

  9. Agent

    Lore runs Lore's own go-to-market

    We dogfood the product to run our company. When a new waitlist signup lands, an agent reads our onboarding runbook from Lore's own brain, researches the company, and drafts personalized outreach that cites the runbook. Our marketing posts work the same way. We review and approve; the agent never sends on its own. Same MCP a customer connects to.

  10. Backend

    Notion in the brain

    Fourth source live. The brain now reads our Notion workspace alongside Drive, Gmail, and Slack. Policy pages, meeting notes, and writeups all flow in and get treated the way each kind deserves.

    Gmail
    connected
    watching 4 labels
    INBOXinvestorsdesign-partnersyc-w26
    last sync 53s ago · +1 thread · +1 fact
  11. Editorial

    Approve a gap, the brain learns

    When the agent runs into something it doesn't know yet, the dashboard now surfaces the gap with the agent's best-guess answer. One click to approve, and next time anyone asks, the answer is there, cited.

  12. Backend

    Policies and memos stay intact

    A refund policy doesn't get useful by being chopped into three context-less sentences. Policy docs, procedures, and memos now land in the brain whole, full text preserved, retrievable as one unit. Meeting notes still get broken down where that's the right move.

    Sources · 3
    May 7obs:7711d4bcPricing memo · "$8/seat for design partners; revisit at 50 seats."›
    May 6obs:a921ee03Re: YC interview prep · "Sankritya, Kinjal, Mohan on Friday."›
    May 5obs:c8d44e21#design-partners · "Heard back from Vercel; thread Friday."›
  13. Backend

    Slack as a real source

    Third source live. Channels you pick get read with their thread structure intact, so a question and its answer stay together instead of landing as two unrelated facts an hour apart.

    Connectors
    poll every 60s
    Google Drive
    active · last poll 47s ago · +2 docs
    Gmail
    active · last poll 53s ago · +1 thread
  14. Agent

    Streaming agent with real tool calls

    Cmd+K opens a chat that picks tools (connector status, last ingest, fact growth, search), runs them in front of you, and streams a markdown answer with cited facts. Works across the whole tenant brain.

    When was the last data extracted?
    ✓last_ingest_event → engine.extracted 3m ago
    ✓recent_facts → 5 facts
    Last extraction was 3 minutes ago from a Gmail thread about the YC interview prepobs:7711d4bc
  15. Backend

    Daily seeder for the Lore brain

    A pg_cron job fires every 5 hours, generates plausible Drive docs and Gmail threads about applying to YC, talking to design partners, and shipping product, and routes them through the same engine pipeline real connectors use. The brain visibly grows between visits.

    today · every 5h
    12:00→ingested 3 docs from Drive · 5 facts
    07:00→ingested 2 threads from Gmail · 4 facts
  16. Editorial

    Editorial narrative replaces the bland header

    The overview now opens with a serif italic lede - "Today, your lore learned 4 facts from Gmail and Drive" - instead of generic KPIs. Cold-start, idle-day, and active-day each get distinct copy.

    today at your lore
    Today, your lore learned 4 facts from Drive and Gmail,
    mostly about pricing for design partners and the YC interview prep.
  17. Charts

    Apache ECharts dashboards

    Replaced the toy SVG sparklines with real charts: a 12-month calendar heatmap, a stacked-area stream graph by provider, a provider→skill sankey, and a force-directed entity graph on /people. Click a calendar day to drill into the timeline.

    facts/day
    by provider
    drivegmailslack
    provider → skill
  18. Editorial

    Sources section finally readable

    Skill bibliographies used to be a wall of opaque [obs:xxx] pills. Now each entry collapses into a one-line preview and expands on click - provider, date, snippet. Source chips next to facts now show the brand logo plus the provider name.

    Sources · 3
    May 7obs:7711d4bcPricing memo · "$8/seat for design partners; revisit at 50 seats."›
    May 6obs:a921ee03Re: YC interview prep · "Sankritya, Kinjal, Mohan on Friday."›
    May 5obs:c8d44e21#design-partners · "Heard back from Vercel; thread Friday."›
  19. Brief

    Brief is temporally aware

    When a fact says "call this Thursday", the brief now resolves it against the email's date and TODAY, drops the event from Next if it already passed, and never reuses relative phrases verbatim.

    today · Sat May 9
    Next, your week.
    Mon May 11Send investor update
    Thu May 14Call Vercel design partner← resolved from "this Thursday"
    Fri May 15YC interview prep with team
  20. Brand

    Cmd+K Ask Lore everywhere

    Press ⌘K from any dashboard page. A pill at the bottom of the screen surfaces it. Submitting routes you straight to the chat with the question pre-filled and auto-sent.

    ⌘KAsk anything…
    »What changed in the brain today?
    »Which design partners haven't responded?
    »Summarize this week's pricing decisions
    »What did Mohan say about the YC pitch?
  21. MCP

    Tenant-scoped MCP server

    Each tenant gets a dedicated MCP endpoint that Claude Code, Cursor, ChatGPT Desktop, and Codex can connect to with a long-lived `lore_…` token. Skills + facts stream into the agent from the brain.

    ~/projects/lore
    $ claude mcp add lore \\
    --transport http https://jointhelore.com/api/mcp/v1/lore \\
    --header "Authorization: Bearer lore_d3…"
    ✓connected · 6 tools available · 3 skills loaded
  22. Backend

    Per-tenant API keys with mint + revoke

    Settings → API Keys lets admins mint a `lore_…` bearer (plaintext shown once, sha256-stored), see prefixes + last-used timestamps, and revoke individual keys. The /start/done page now has a one-click test button that probes the live MCP server with the new key.

    API Keys
    + Mint key
    Production agent
    lore_d334c7b… · created 1d ago · last used 12m ago
    Revoke
    Staging probe
    lore_a921ee0… · created 4d ago · last used 2h ago
    Revoke
  23. Onboarding

    Live ingest with per-file progress

    /start/ingest now shows what the engine is reading right now - file titles landing, fact counts, threshold-aware status ("3 more facts until your first skill compiles"). Auto-advances at 1 skill or 5 facts.

    Reading 3 of 7 docs…
    3 more facts → first skill
    pricing-memo.md+2 facts
    Re: YC interview prep+1 fact
    design-partner-list.mdreading…
    kpis-april.mdqueued
  24. Backend

    Skill auto-emit cron

    A pg_cron job every 5 minutes scans every workspace, finds skills missing from the file system or with ≥3 new facts since their last emit, and recompiles the highest-priority one. No more manual emit calls.

    skills/design-partner-outreach.md
    recompiled 2m ago
    name:design-partner-outreach
    version:1.0.4
    sources:drive, gmail
    observations:11
    When this Skill applies
    A teammate asks how to approach a new design partner…
  25. Backend

    Connector polling cron

    Vercel Cron + pg_cron pull every minute, check each tenant's connectors for new items, fetch them, and hand each to the engine. The first version of the live data path: every Drive doc / Gmail thread you write shows up in the brain within ~60 seconds.

    Connectors
    poll every 60s
    Google Drive
    active · last poll 47s ago · +2 docs
    Gmail
    active · last poll 53s ago · +1 thread
  26. Backend

    Gmail connector

    The second provider through the same machinery: OAuth, watch-list of labels/threads, fact extraction with provider chips. Identical reliability to Drive - just a different shape of input.

    Gmail
    connected
    watching 4 labels
    INBOXinvestorsdesign-partnersyc-w26
    last sync 53s ago · +1 thread · +1 fact
  27. Onboarding

    Editorial /start onboarding

    Five-step flow with a serif italic lede on each page: workspace → connect → pick → ingest → plug-in. Each step is one decision, not a checkbox grid.

    step 3 of 5
    Pick what your lore should learn first.
    1workspace
    2connect
    3pick
    4ingest
    5plug-in

Want in early?

Lore Pilot is onboarding its first cohort from the waitlist. Lore Fleet is opening to a small group of design partners, teams with a few years of Slack / Drive / GitHub history who want their agents to inherit it. Tell us which one sounds like you.