What we've shipped

ELAV now checks new facts against existing Brain context, filters stale facts from answers, refreshes old briefs in the background, and keeps entity recall tied to the right source timing.

• Contradictions are flagged before they become trusted Brain memory.
• Ask can recall people and companies through Brain candidates instead of loose text matches.
• Old or orphaned entity context is refreshed and reconciled automatically.

Home now tells you when connected sources are still being read, preserves useful briefs when context already exists, and brings urgent attention signals into Ask.

• Cold-start scan progress is based on completed backfills, not stuck first runs.
• Existing accounts no longer lose their brief to a generic reading state during sync.
• Attention cards carry urgency and source provenance into Ask context.

Source OAuth and Brain reads now fail closed on data residency, source provenance, and permission checks, with diagnostics to catch stale access before answers rely on it.

Zoho Mail and Zoho Calendar are now available as sources, so ELAV can read recent mail and calendar context from Zoho accounts alongside Gmail, Slack, and Google Calendar.

Admins can now create, copy, and revoke Brain MCP keys from Settings, with token-safe setup guides for Codex, Claude, OpenClaw, and generic MCP clients.

Workspace creation is more reliable, active sessions route to the right place, and personal email sign-ins are blocked earlier across email and Google OAuth — except for the approved founder account.

Set up a workspace during onboarding and invite teammates by email. Everyone sees ELAV through their own permissions — what they can already access in connected tools is what they can use here, with no manual access setup.

• Invite and remove members, with clear Owner, Admin, Member, and Viewer roles.
• Removing someone revokes their access immediately.
• No access matrices to manage — source permissions carry over automatically.

Each person and company brief now shows how they connect to the rest of your work, a timeline of sourced context, and a private note only you can see.

• A relationship map surfaces the most relevant connected people and companies for each brief.
• Add a private note to any person or company that stays visible only to you.
• Every brief keeps source receipts and timing so you can trace where context came from.

Brain Memory now groups what changed in your company context and lets you rewind any fact to see its before and after, with source receipts and timing kept intact.

• Grouped change summaries make it easy to scan what's new without digging through sources.
• Per-fact rewind shows the earlier and current state side by side.
• Private source-backed context stays filtered to the people allowed to see it.

Older context that was previously held back now shows up in Memory and answers, still filtered to who is allowed to see each source — including refreshed Slack channel access.

ELAV now re-verifies open commitments on a schedule, so promises and follow-ups across your sources do not quietly slip through.

Ask Brain now works from a single entry point and adapts to where you are asking from — a person's brief, your workspace, or the main Ask page — so answers stay grounded in the right context.

Meeting prep and morning briefings can use relevance-scored Brain facts and safer contact context, so important work history is less likely to get buried by noise.

• Admin curation controls let each workspace define what counts as useful context.
• Contact dossiers preserve source receipts while giving meeting prep and briefings better people context.
• New calibration and smoke checks make it easier to validate relevance quality before relying on it.

The Inbox review flow now separates urgent approvals, captured tasks, memory proposals, and quiet activity into a simpler daily review surface.

• Memory proposals show clearer summaries and a Brain safety note before review.
• Captured tasks are easier to scan without the old dense table feel.
• Quiet activity now reads as one review stream instead of a grid of competing cards.

The website now explains ELAV around company memory, sourced answers, reviewed work, source maps, and agent memory with tighter proof and fewer duplicate sections.

• Homepage sections were rebuilt around sourced answers, source maps, history, and agent/tool memory.
• Pricing CTAs, FAQ, footer copy, and social preview metadata now match the new positioning.
• Product branding is cleaner too: source logos, the browser app icon, and first-run connection cards now use ELAV's shared logo system.

Home, Brain, Sources, Settings, and Ask now present the everyday ELAV workflow more directly, with less internal structure showing through.

• Brain Sources now owns source setup, including Meeting Intelligence, while legacy Connections flows move out of the main path.
• Home and Ask defaults avoid internal/test wording and keep first-run guidance focused on connected work.
• Settings, sidebar workspace links, member management, and Brain entity surfaces were simplified for repeated daily use.

The Brain Memory feed now brings together important updates and superseded facts, so you can see what changed in company context without digging through old source material.

• Memory shows when a fact stopped being current, not only when it was first learned.
• Each change keeps source receipts and timing so the feed is easier to audit.
• Private source-backed context stays filtered to the people who are allowed to see it.

Connected AI agents now get a safer read-only view of Brain capabilities and source readiness before they answer with your company context.

• Agent setup now starts from memory capabilities and source coverage instead of obsolete campaign tools.
• Source coverage shows freshness and readiness without exposing raw source records or credentials.
• Brain MCP access stays scoped to the credential you issue, with read-only tools for cited evidence and timelines.

The public site is sharper, the product now has its first Home brief, and the Brain write path is safer under load.

• The new website explains client memory with clearer proof, pricing, booking, and Ask examples.
• Home now starts with a brief: next meeting, today strip, attention cards, and first-run connection guidance.
• Brain writes and Ask processing are steadier, with quieter progress states and safer backlog handling.

Open any Inbox item into a focused Details page with its status, activity timeline, and source thread side by side — replacing the older Operator view with something quieter and easier to scan before approving.

If a question needs you to reconnect Gmail, Slack, or Calendar mid-thread, Ask now resumes the original prompt automatically once the source is back — no need to retype or refresh.

Webhook triggers now show "Not used" when a source doesn't need them, and a failed trigger install only flags a connection as "Needs attention" when the Brain hasn't actually learned anything yet — so healthy sources stop nagging.

New docs cover connecting sources, asking with evidence, reviewing Inbox outcomes, approvals, source sync, privacy, and troubleshooting.

Ask can pull recent context from connected sources, turn it into cited digests, and keep the final answer saved with the conversation.

Connections now handle custom Slack setup, live sync state, and learned evidence so healthy sources are not mistaken for sources that need a refresh.

Source sync now reflects fresh activity more accurately, recovers from ingestion delays, and keeps connection status aligned with what the Brain has learned.

Ask, Inbox, and Details now keep richer progress, evidence, and retry state together so generated work is easier to inspect before approval.

Ask now has clearer empty, loading, and evidence states so you can tell what ELAV is ready to do before a source-backed answer begins.

Timeline-style answers now preserve provider event times and rank evidence more conservatively, so dated answers stay closer to what happened when.

Workspace tools now scope available source context to the accounts you connected and provide clearer artifact review controls.

The workspace now keeps requests, source evidence, generated artifacts, and approval drafts together so you can inspect what ELAV found before anything is sent or changed.

Ask now searches your connected sources for follow-up questions instead of demanding clarification. When nothing's found, it says so honestly without overclaiming.

Connected sources now show the email or workspace they're linked to — Gmail, Google Calendar, and Slack — so you can tell at a glance which account is wired up.

Surfaces inbound asks across your recent Gmail threads in one view. Read-only — nothing acts without your approval.

The Inbox is now a sortable table. Every task has its own URL, openable from the table or a side drawer.

Capture a task from any Slack message. ELAV picks out the actionable ones; approved tasks land in your Inbox, linked to the source.

We now track every action ELAV takes for you — answers, suggestions, approvals, tasks. Ask for a weekly summary of what shipped. No raw inbox content stored.

The Brain dashboard shows the people, companies, and recent context it's actually learned from your sources.

Brain has its own workspace now — Overview, Sources, Memory. Source status reads as Ready, Learning now, or Needs refresh, instead of admin output.

Answers from ELAV now respect each source's permissions. Private emails, restricted calendar events, and shared docs only surface for people who already had access.

A new panel on Brain shows the status of every connected source. Sync issues now surface clearly instead of failing silently.

Long answers stream cleanly to the end and stay readable when saved — no more truncation or flicker.

Type a request the way you'd say it. ELAV picks the right skill and runs it.

Newsletters, automated alerts, and admin notifications no longer leak into ELAV's suggestions. Skills stay grounded in the work that matters.

Slack, Gmail, and Calendar connections are scoped to you — not your team's workspace.

Conversation history, an Inbox of pending approvals, and a Brain dashboard — all in one place.

A searchable drawer over thousands of tools, grouped by Email, Calendar, Communication, and Documents. If something isn't supported yet, request it in one click.

Type a question in the sidebar. ELAV decides whether to search your Brain or run a skill, and answers.

ELAV is live. Connect Gmail and your Brain starts learning your team's work — surfaced through a real-time dashboard, with approvals required before anything acts. Self-hosted in the EU.