VOICE-FIRST · LOCAL-FIRST · OPEN SOURCE
Say it. Consider it handled.
From conversation to completion — it calls you back when it's ready for your review.
SayDo is a voice-first chief of staff for your projects. Just talk it through — it studies your project first, keeps a four-color ledger of everything, and once it truly understands, it drives the AI tools already on your machine to do the work. When the run and its checks are done, it calls you back for review.
Desktop service · free & open source iOS / Android / HarmonyOS · Coming soon
Today
TodayA complete loop
You talk and you decide; it does the heavy lifting. When the run and its checks are done, it calls you back for review — nothing counts as delivered until you sign off.
Just talk
Hold to talk, think out loud. A casual "note this for me" lands in the ledger instantly. With your project's context loaded, it interviews you like a good podcast host — one question at a time.
Readiness check
It decides when it knows enough and proposes to start — instead of asking you to write a spec up front. The spec is its output, not your input.
Decision pack
Three things land in front of you: an outcome preview (what you'll get), a plan (each step marked AI-run or needs-you), and a lightweight demo of the end state.
You decide
Look at the preview and say "go". Merging and delivering are always your call — never automatic.
Background execution
Runs on your own machine: isolated git worktrees leave your main workspace untouched, with budget, time, and turn circuit-breakers. You can walk away.
It calls you back
When the run and its checks finish, it says "ready for your review" — not "done". Escalation runs from in-app voice callbacks to desktop notifications; mobile push arrives with the apps (coming soon).
Review & settle
A spoken summary plus an evidence view (diffs, tests, decisions), organized against your acceptance criteria. You approve, the entry is stamped settled, and what it learned stays for next time.
Things said are forgotten; things promised are nowhere to be found. Who's waiting on whom, what's in flight, how it's going — all kept in your head and scattered across chat threads.
Everything has an owner and a state, visible at a glance: awaiting your call, owed by you, in the AI's hands, or waiting on the outside world. Four colors, nothing lost.
Not another voice tool
Every piece of the puzzle has a mature point solution. Nobody has closed the full loop — talk it through, start itself, deliver for review — under one governance model. SayDo does the whole loop.
Said is booked
Hold to talk, think out loud. A casual "note this" is booked instantly — no app to open, no wording to polish.
It learns before it works
The first conversation about a project starts with deep research, distilled into a persistent knowledge base. That's the line between SayDo and a fresh chat window.
Real AI, doing real work
It drives the AI tools already on your machine — Claude Code, Codex, Cursor, Gemini CLI, and more coming. Whichever you subscribe to, it uses. No extra fees.
You stay the decider
Work starts only after a decision pack gets your sign-off, and finishes with "ready for your review" — never "done". Merging and delivering are always your call.
Your data stays yours
No cloud servers, no accounts. Your phone pairs with your own computer over your LAN. The developer cannot access — and does not collect — any of your data.
One ledger across devices
What you say on your phone shows up on your computer; what's booked on your computer is watchable from your phone. Communication is mobile; execution stays put.
Where things stand, honestly
No vaporware. What's usable today says "Available now"; what's still on the road says Coming soon.
Desktop service
Available nowThe execution side on macOS: daemon, voice pipeline, and web console. Free and open source — your own computer is all the infrastructure there is.
Drives the AI you have
Available nowPlugs into signed-in AI tools on your machine — Claude Code, Codex, Cursor, Gemini CLI, and more. Whichever subscription you have is the one it uses.
Memory & the ledger
Available nowProject grounding, a persistent knowledge base, the four-color ownership ledger, and evidence views for review. What you discuss settles in — it knows you better over time.
iOS app
Coming soonThe communication side in your pocket: hold to talk, watch the ledger, approve decisions. Working toward store release.
Android app
Coming soonThe same communication surface as iOS, connected to your desktop service over your LAN.
HarmonyOS app
Coming soonA HarmonyOS communication surface, on the roadmap.
Call-style voice reports
Coming soonWhen your computer finishes a run, your phone rings like an incoming call — pick up and it briefs you. PushKit wake-up with a CallKit call UI.
Talk on your phone, work on your computer
Communication is naturally mobile; execution is naturally stationary. Your phone only talks — voice, progress, approvals. The heavy lifting stays on your computer.
Communication · your phone
Talk, watch progress, decide, approve. Always at hand — and if it disconnects, the backend doesn't care.
Execution · your computer
Studies the project, drives the AI, works in the background, keeps the memory. Step away from your phone — execution goes on.
No servers. Period.
Your conversations, voice, items, and ledger live on your own devices and travel only across your own LAN. No accounts, no analytics, no third-party tracking or ad SDKs. The developer cannot access — and does not collect — any of your data.
Read the full privacy policyOpen-source desktop, apps on the way
The desktop service is the execution side — free and open source. The mobile apps are the communication side, working toward store release. Run the desktop service on your macOS, scan the QR code with your phone, and you're connected. No API keys to apply for — any signed-in AI tool on your computer is enough to start talking.
What it is not, first
Is this a voice keyboard?
Does it upload my project to a cloud runner?
What do I need to start?
Where is my data stored?
How does it know when something is finished?
Connect the saying to the doing
The desktop service is free and open source, and the apps are on the way. From today, everything you say has a place — and a fate.