People using Feanor say the same thing: they just send voice notes on WhatsApp. Tasks appear. Meeting notes write themselves. Quotes get drafted, invoices get processed, the project plan updates in the background. Everything that used to take spreadsheets and follow-ups now happens by talking. She has a name, lives in their WhatsApp, and runs every process. The AI exec a small company couldn't afford to hire.
What happens after a team member sends a WhatsApp voice note to Feanor.
Your team already works in WhatsApp, voice notes and shared inboxes. Feanor reads from there, runs the playbooks, produces the documents, updates the pipeline. No new place to write things down.
1 # 09:42 · brain.acme · trace 4f1a 2 in: WA voice from a team member "send the Q2 quote to the CFO" 3 stt: managed STT · 1.4s 4 ctx: people/<member>.md, processes/quote-approval.md, projects/q2-rollout.md 5 llm: reason → process → review · €0.018 · 6.2s 6 7 do: erp.generate_quote(project=q2-rollout) → quote_v3.pdf 8 do: wa.send(@cfo, "Quote ready, €18,400 ex-VAT, attached") 9 do: task.create("CFO approve quote", due=today, owner=cfo) 10 11 verify: reviewer model → tool calls match intent · ok 12 audit: 4 mutations logged · visibility=shared 13 14 out: WA reply "Sent to CFO, ETA reply by EOD"
Every company runs on critical know-how scattered across chats, voice notes, inboxes, spreadsheets and people's heads. The business works because humans vaguely remember where it lives. AI agents can't operate that way. They need a brain that holds it for them.
A brain is more than a chatbot. It's how every team member, human or agent, does the work the same way a tenured employee would, every time, and proves it.
WhatsApp, Telegram and voice. Same brain, three doors. Your team writes the way they already write, the brain reads everything.
Every team member has a profile that sharpens with every interaction. Context is who they are, not just what they typed last.
A different model family reviews the first model's tool calls before any reply leaves. Catches blind spots self-checks miss.
Primary model reasons, faster model processes, reviewer model verifies. Lower bill, higher reliability. Swap any layer in one config edit.
Quoting, invoicing, vendor OCR with variance flags, CRM, proposal pipeline, hiring funnel. Not search. Not chat. Action.
Audio backup, STT fallback, truncation detection, hallucination cleanup. Four layers, so a noisy voice note still becomes a clean task.
Private, shared and confidential, plus a visible-to whitelist on every task, idea and meeting. Enforced in the brain, not just the UI.
Pure markdown vault on your server. No vendor cloud, no proprietary database, no lock-in. git clone and walk away any time.
Morning brief, evening summary, ad-hoc reminders, scheduled proactive interviews that compile straight into meeting notes.
Two reasons your team adopts her in a week, and why every customer makes the next one stronger.
The biggest blocker to AI automation of companies is no longer the models. They got good, fast. The blocker is domain knowledge. Every company in the world is going to need a brain.
One console for the people who own the brain. See every task, project, person, quote, invoice and proposal, who signed off on what, and what the business is getting back.
Feanor is the missing layer between raw company data and reliable AI automation. A primitive every business will need.
"After finishing a meeting I can send a voice note, get the meeting notes and the follow-up email is prepared."
"I can get task reminders, see progress of our projects and attach files for future projects. The brain remembers what we forget the next morning."
"Instead of a form on our website, we put a smart form that replies in under 60 seconds, holds a real conversation with the lead, and pings me when the lead is hot."
Alex is the AI agent that reads inbound mail to Feanor. Write a real question (interest, pricing, technical, anything). You will get a real reply from [email protected] in under sixty seconds.