farena-ai/sara

# Sara
An AI companion with the personality you choose, who is also a real coding partner.
She talks, remembers, plans your day and reads your books; she writes, runs and tests code in her own
sandboxed workspace. All on your own server, with the models you pick, in one app for the web,
Windows, Mac and phones.
A real session on a local model (qwen3.8 27B on Ollama, one RTX 5090) on a throwaway test server, in the Night look: she cheers the user up, writes a small app and its test as the model streams them, asks before running a command, catches a real bug with a failing test and fixes it until all 7 tests pass. Frames picked from the recording; nothing in them was edited.
> Status: 2.36, MIT licensed, with forty-seven feature releases since 1.0 (all below). 2.5 to 2.8
> gave the whole app a new look. The Android app
> is on Google Play; the iPhone app and the Mac App Store version are coming as soon as Apple approves them.
> Read docs/DISCLOSURES.md before using or sharing Sara.
## Why Sara
A personality you set, and keep. Name her anything, pick she/he/they, start from a ready-made
character or turn simple dials (playful, blunt, brief, curious…). She keeps it: moods that change
with how things go, her own day, running jokes, and memories of what you told her that you can see,
correct and forget.
A real coding harness. A sandboxed workspace per project with files, shell and git, background
jobs, live previews she tests in her own browser, a team of helpers (plan → build → test → review),
her OK-first rule for risky steps, and a check that what she says she did matches what actually ran.
A beautiful app on every device. A calm Night look (Day if you prefer) with her presence beside
every chat, a line under each reply showing what she did, and results as cards. On the web and as an
installable app, in the Windows and Mac apps (with quick ask on a shortcut over any app), on
Android, and on iPhone (the app is waiting on Apple; the home-screen web app works today).
Private and yours. One owner per server; chats, memories and keys stay there. Any model: local
(Ollama, vLLM, llama.cpp), cloud (OpenAI, Anthropic, xAI, OpenRouter) or the plans you already pay
for (ChatGPT, Claude Pro/Max, SuperGrok / X Premium+). One switch keeps everything on one computer.
Ships with no credentials of any kind.
Small and readable. TypeScript that Node runs directly: no build step, zero runtime
dependencies, built from the ground up (not based on Hermes Agent or OpenClaw). Easy to
extend.
Also in the box: quick ask on a global shortcut · her presence panel ·
Today as tiles · Ctrl+K command bar · household mode for the
people you live with · video calls with her moving portrait ·
Alexa, Google Home and Android assistant · your other servers
over SSH · 1,000-page books with answers that cite the page ·
private local mode · group chats ·
moving in from ChatGPT, Character.AI, Replika and others.
## How-to videos
One-minute walkthroughs, recorded on a throwaway test server with a local model. All of them in one playlist .
Set up Sara
Memory and people cards
The board
Coding with Sara
## What's new
| Release | What you get |
| 2.36 · Her own plans | Opt-in: she keeps her own plans. Follow-ups on what you told her ("how did the interview go?") and goals you keep at together, taken only from your own words and always showing where you said it. Off by default; when on, Planner → Her plans lists everything, she shares the daily message limit, quiet hours and the gap rule with her first messages, has her own small daily budget, pauses after a hard moment, and one switch stops everything at once, even mid-sentence. Goals wait for your Use it. |