HAL

The meeting agent
for people who'd rather
build than attend.

Hal autonomously joins your Google Meet, Zoom, and Microsoft Teams calls on your behalf — listens, takes notes, speaks when you want it to, and follows up after. Your tokens, your transcripts, your infra. Open source.

  • Connects to Google Calendar and Microsoft 365 and decides which meetings deserve Hal.
  • Joins Meet, Zoom, and Teams via official SDKs and a hardened headless browser.
  • Listens, takes notes, and speaks on your behalf with disclosed bot-as-delegate identity.
  • Drafts follow-up emails, updates your CRM, and books the next step — for your review.
meet.google.com · launch-review · 11:00 AM
Maya Chen
Devon Park
HAL · AIfor Halleluyah Oludele
Priya R.
Live transcriptrecording · disclosed
  • Maya…so the launch date slipped a week, but QA signed off this morning.
  • Hal (AI · for Halleluyah Oludele)Noted. I'll move the release announcement to next Friday and update the launch doc.
  • DevonPerfect. Can someone also loop in support before then?
  • Hal (AI · for Halleluyah Oludele)Drafting an email to support@ now — I'll send it for review after the call.
  • Hal is listening…
A day in the queue

Hal in your week.
Four meetings, four modes.

Some meetings need you. Some need a summary. Some need someone to speak on your behalf. Some shouldn't happen at all. Hal sorts the difference, on a per-meeting policy you set.

09:00

Engineering standup

skipped

You didn't go. Hal listened, took notes, emailed you a 90-second summary at 09:13. Your team didn't notice you weren't there.

Skipped · summary emailed
"Hal: '3 PRs merged, infra deploy at noon, no blockers raised.'"
auto-cycling · hover to pause
How it works

A meeting agent that onboards itself
and starts taking calls on day one.

May 19

Hal's onboarding checklist

~12 min

Hal makes itself part of your team

  • Connect Google Calendar or Microsoft 365
  • Connect Google Meet, Zoom, and Teams
  • Upload your meeting persona (optional)
  • Configure per-meeting policy: auto, ask, ignore
  • Connect your CRM (HubSpot, Salesforce, Attio)
  • Connect Notion / Linear / Slack for follow-ups
  • Add an integration (Hal helps you pick one)
What to expect

Hal's first week

  • Day 0Get started

    Hal indexes the next 7 days of meetings. You set policies for each one.

  • Day 1–3Listen-only

    Hal joins as a disclosed participant. Listens, transcribes, emails you summaries.

  • Day 4–6Get comfortable

    Enable chat-on-behalf in low-stakes meetings. Hal answers logistics in the chat for you.

  • Day 7+Speak mode

    Toggle voice-on-behalf for meetings you trust Hal with. Always disclosed. Kill switch always on.

Manifesto · 01

"We don't think every meeting deserves a human. We think most deserve a summary, a few deserve a delegate, and almost none deserve another hour of your week. Hal is the agent that sorts the difference — on the record, with your tokens, on your infra."

H
Halleluyah Darasimi OludeleBuilt Hal · hallelx2 · 2026
Waitlist · pre-alpha conversations

Why people are signing up.

Hal is pre-alpha — none of these are customers yet. They're paraphrased from conversations with people on the waitlist about why they want Hal in their week. Click a name on the left, or wait six seconds.

"Half my calendar is meetings I shouldn't be in. The other half is meetings I shouldn't be running. Hal handles the first half and I get to think about the second."

TA
Tomi A.Engineering Manager · fintech, lagos
pre-alpha
Built for
FoundersEngineering teamsDevRel orgsSolo devsCustomer successSales engineersProduct leadsRemote teamsOSS maintainersConsultantsVC firmsCompliance-bound enterprisesPeople with 30+ meetings a weekPeople who skip standups on purposeFoundersEngineering teamsDevRel orgsSolo devsCustomer successSales engineersProduct leadsRemote teamsOSS maintainersConsultantsVC firmsCompliance-bound enterprisesPeople with 30+ meetings a weekPeople who skip standups on purpose
How Hal compares

To the four honest options for skipping the meeting.

Either you attend, you record, you build, or you delegate. Hal is the last column — and the only one that exists as a self-hostable open-source project.

The default
Do nothing
  • Same meetings every week, same drain.
  • No record. No summary. Pure entropy.
  • You miss the call → you miss the context.
Alongside-you bots
Otter / Read AI / Fireflies
  • Joins beside you. Records. Summarizes.
  • Cloud-only, can't self-host. Your transcripts live on their servers.
  • Useless if you can't actually be in the meeting.
The Saturday plan
Build it yourself
  • Three platforms. Headless Chromium. SDK approval queues. KMS.
  • Six weekends in, you have a Meet bot that almost works.
  • You're now in the meeting-bot business, not yours.
The agent that goes for you
Hal
  • Joins for you when you can't. Speaks on your behalf, disclosed.
  • Self-host on a VPS or use the eventual hosted plan. Same code.
  • Encrypted by default, BYO STT/TTS/LLM. AGPL-3.0.
Get early access →
* Hal is pre-alpha. The comparison reflects the v1.0 product Hal is aiming at, not what's shipped today.
Capabilities

Hal handles meetings
like a high-agency teammate.

Hal isn't a transcription tool with a chat. It's an agent with a role, a voice, a kill switch, and an audit log. Built to be trusted in the room.

Joins for you

When you can't attend, Hal joins as a disclosed delegate. Bot-as-delegate, never deepfake.

Listens like staff

Speaker-diarized transcripts. Action items, decisions, open questions surfaced automatically.

Speaks when you want

Pre-brief Hal on goals, talking points, hard nos. Confidence threshold before speaking.

Drafts for review

Follow-up emails, CRM updates, calendar holds — all drafted, never auto-sent.

Encrypted by default

AES-256 envelope encryption, per-user DEK, KMS-wrapped. Your tokens never leave plaintext.

Bring your own everything

Whisper or Deepgram. Piper or ElevenLabs. Ollama or Claude. Local libsodium or AWS KMS. All swappable.

Easy to take over

Dial in mid-meeting and Hal mutes itself. Every line Hal said is in the audit log.

Open source, AGPL-3.0

Fork it. Run it. Contribute back. Self-host on a VPS or wait for the hosted plan.

The four questions

Every meeting agent answers these.
Hal answers them out loud.

01Question 01 / 04

How does Hal enter the meeting?

Native SDKs where they exist — Zoom Meeting SDK, Microsoft Teams Bot Framework. A hardened headless Chromium worker for Google Meet (because Meet has no production bot API). Whatever the platform, Hal joins like a participant, on a real connection, with real audio.

Zoom Meeting SDKTeams Bot FrameworkHeadless Chromium
HeadlessMeet
SDKZoom
SDKTeams
02Question 02 / 04

Who does Hal speak as?

Bot-as-delegate. Never deepfake. Hal joins as "Hal · AI for <your name>" and announces itself on entry. Voice cloning is opt-in, requires a recorded consent step, and every generated clip is watermarked and audit-logged. You can always disable speaking entirely.

DisclosedConsent-gatedAudit-loggedKill switch
Identitydisclosed
Display name
HAL · AI
for Halleluyah Oludele
  • Announces on join
  • Voice clone opt-in only
  • Every utterance audit-logged
03Question 03 / 04

When does Hal show up?

Connect Google Calendar or Microsoft 365 once. For each meeting, set policy — auto-join, ask first, or ignore. Hal subscribes to your calendar via push, so when a 10:30 sync becomes a 10:45 sync, Hal updates. Recurring meetings inherit your policy.

Calendar pushPer-meeting policyRecurring inheritanceOn-time joins
Tomorrow4 events
  • 09:00StandupIgnore
  • 10:30Launch reviewAuto-join
  • 13:00Design syncAsk first
  • 15:30Investor updateAuto-join
04Question 04 / 04

What does Hal do after the call?

Transcript. Summary. Action items. Drafted follow-up email (for your review, not auto-sent). Optional CRM updates, Notion handoffs, calendar holds for promised next steps. Hal is opinionated about not pretending decisions were made when they weren't.

TranscriptsDrafted follow-upsCRM + NotionAudit trail
After the meetingdrafted
  • EmailFollow-up to support@ · ready for review
  • CRMUpdated Acme deal stage to "Proposal sent"
  • CalHeld 15 min Friday for launch announcement
  • NotionLogged 4 decisions, 2 open questions
Envelope encryption
KMS · Master Key (KEK)never leaves KMS
Per-user DEK (AES-256-GCM)stored as ciphertext
Tokens · Transcripts · Audiociphertext at rest, in transit, in logs
What this protects · what it doesn't
  • A stolen DB dump
  • A leaky log line
  • A malicious operator with read-only DB access
  • × Shell access to a bot worker mid-meeting
  • × A compromised KMS master key
Encrypted by default

Your tokens never leave plaintext.
Not even once.

Hal uses standard envelope encryption: a master key in your KMS wraps a per-user data encryption key, which encrypts every OAuth token, transcript, and audio file. Lose the database — lose nothing.

  • AES-256-GCM
    Authenticated encryption everywhere
  • Per-user DEK
    One breach can't fan out
  • Pluggable KMS
    Vault · AWS KMS · GCP KMS · local
  • Zero plaintext logs
    Tokens redacted at the logger
  • Bot-as-delegate
    Always disclosed, never deepfake
  • Audit trail
    Every action Hal took, in writing
Self-host

Your infra.
Your rules.

One docker compose up on a VPS and you have your own private meeting agent. No phone-home, no required SaaS dependency, no telemetry. The same codebase will eventually power the managed hosted plan — but you're not required to wait.

ProviderLocalCloud
  • Speech-to-textWhisper
    DeepgramAssemblyAI
  • Text-to-speechPiper
    ElevenLabs
  • LLMOllama
    AnthropicOpenAI
  • Key managementlibsodium
    Vault TransitAWS KMSGCP KMS
  • Object storageLocal FS
    S3GCS
~/hal-meetings
$ git clone https://github.com/hallelx2/hal-meetings
$ cd hal-meetings && cp .env.example .env
$ docker compose up -d

→ web     listening on http://localhost:3000
→ agent   bot worker ready
→ vault   transit engine initialized
→ db      postgres up · migrations applied

first run  open http://localhost:3000/setup
# OAuth your calendar. OAuth Meet/Zoom/Teams.
# Drop a meeting URL. Watch Hal join.
Phase 0 ships in < 5 min on a 2-core VPS
Roadmap · pre-alpha honest

What's shipped. What's next.
No mystery boxes.

Hal sequences by the hardest engineering problem first — joining meetings reliably — not by the most demoable feature. ETAs are best-effort estimates from a one-person team. They will slip. The roadmap will not.

P0

Meet · listen-only

Headless Chromium bot joins Google Meet as a disclosed participant. Transcribes. Emails you a summary. Manually triggered with a URL.

In progress
P1

Calendar auto-join

Google Calendar + Microsoft 365. Per-meeting policy: auto, ask, ignore. Hal joins on time, every time.

Up next
P2

Chat on behalf

Hal posts in the meeting chat for you — answers logistics, drops links, surfaces docs. Controlled action surface before audio.

Planned
P3

Speak on behalf

Voice mode, opt-in. Pre-meeting brief. Confidence threshold before speaking. Kill switch. Audit log of every utterance.

Planned
P4

Zoom SDK worker

Native Zoom Meeting SDK worker for higher-quality audio and stable join behavior. Marketplace submission.

Planned
P5

Teams SDK worker

Microsoft Teams Bot Framework integration. Calling and meeting bot. Partner Center submission.

Planned
P6

Post-meeting action layer

CRM updates (HubSpot, Salesforce, Attio). Notion / Linear handoffs. Calendar holds. All drafted, never auto-sent.

Planned
FAQ

Questions you may ask (we get these frequently)

  • +What is Hal, in one sentence?
    An autonomous, self-hostable meeting agent that joins Meet, Zoom, and Teams on your behalf — listens, optionally speaks, and follows up after.
  • +Is this legal? Doesn't recording need consent?
    Most jurisdictions require notice (one-party) or affirmative consent from every participant (all-party). Hal joins as a named bot ("Hal · AI for <your name>"), announces itself on join, and respects regional policy you configure. In all-party-consent regions, Hal can be set to wait for explicit consent before recording or speaking.
  • +Does Hal pretend to be me? Is this a deepfake product?
    No, and we will not build that mode. Hal is bot-as-delegate. It always appears with "AI" in its display name and discloses itself on joining. Voice cloning is opt-in only, gated behind a consent flow, watermarked, and audit-logged. If you want a deepfake, this is the wrong project.
  • +Where do my transcripts and tokens live?
    On your infrastructure when you self-host. Every OAuth token, transcript, and audio file is encrypted with a per-user data encryption key, which is itself wrapped by a master key in your KMS (Vault Transit, AWS KMS, GCP KMS, or local libsodium for development).
  • +Which platforms work right now?
    Phase 0 — Google Meet only, listen-only, manually triggered. Zoom and Teams (via their official SDKs) and calendar auto-join come next. The roadmap section is honest about ETAs.
  • +Can I run Hal without any external APIs?
    Yes. Pair Whisper for STT, Piper for TTS, Ollama for the LLM, and local libsodium for KMS. Latency suffers and quality is lower than managed services, but no audio or text ever leaves your box. This is the 'fully air-gapped' tier and a first-class supported configuration.
  • +When will hosted Hal be available?
    After the self-host experience is solid. The same codebase will run both modes — managed Hal is just a deployment of the open-source project with cloud KMS, hosted STT/TTS, and billing turned on. No SaaS-only features.
  • +How is Hal different from Read AI, Otter, or Fireflies?
    Those products join alongside you. Hal goes for you — listening when you can't attend, speaking on your behalf within boundaries you set, and following up after. Plus: open source, self-hostable, encrypted by default, and explicitly not a deepfake.
  • +Who built this and why?
    Halleluyah Darasimi Oludele (hallelx2 on GitHub) built Hal because the existing meeting tools either join alongside you or require a 12-step Cursor build. Both feel wrong for the post-2026 "agents that go for you" pattern. The project is solo, open, and AGPL-3.0.
Pre-alpha · invite-only

Send Hal to your
next meeting.

We're rolling out access in small batches. Drop your email and we'll send a setup link when Hal can join a meeting in your stack.

We'll only email about Hal access. No newsletter, no tracking pixels.