Keres keeps work, clients, shopping, and research in completely separate browsers — then automates the boring parts on a schedule. Your logins never mix, your data never leaves your machine.
Each in its own browser, each with its own logins. You changed nothing.
How it works
No browser-engineering degree required. You make simple choices; Keres handles the machinery underneath.
Give it a name — “Client work”, “Shopping”, “Research”. Each profile is its own browser: own cookies, own logins, own history. Nothing leaks between them.
Pick an environment — Standard, Private, Automation — instead of fiddling with fifty settings. Sensible defaults; expert controls stay one click away.
Browse normally, or put the profile to work: record what you do once, replay it daily, and wake up to finished work.
Automation
Show Keres a task once — it records your clicks like a macro, then replays them exactly, daily, weekly, or on the dates you pick.
Live schedules from a real Keres workspace.
Why Keres
Separate storage, cookies, and sessions per profile — enforced by the runtime, not by wishful thinking. One client's login can never bleed into another's.
Schedules, retries, missed-run policies, and full run history. Set the cadence once; Keres handles drift, overlaps, and failures.
Two engines behind one stable interface, and a hardened Camoufox runtime for Private profiles: screen, fonts, canvas, and timezone presented coherently out of the box. Pick per profile; your tasks work identically everywhere.
OpenAI, Claude, Gemini, Groq — or Ollama running on your own machine. Keres shows honest connection status and never pretends an offline server is reachable.
A native Model Context Protocol gateway lets AI assistants drive your profiles through safe, permission-checked tools.
Profiles, history, credentials in your OS vault, backups to your own storage. The cloud is an option for sync — never a requirement to work.
A hardened Firefox-based runtime with built-in identity defenses. Pair it with the Private environment in one click — no fingerprint artisanry required.
Connect a provider once — Keres agents browse, decide, and act inside the right profile.
AI, grounded
Keres speaks the Model Context Protocol natively, so Claude, GPT, and open models can operate your profiles — with permission boundaries, human approval where it matters, and a full audit trail of everything they touched.
Pricing
One automated report a week covers Pro for most people. No seat math, no credits to count.
For trying Keres and everyday browsing.
For professionals who live in their browser.
Questions
No. Keres is a browser with isolated profiles and automation. You can route a profile through your own proxy if you want, but Keres doesn't sell or provide network services — it organizes and automates your browsing.
No. Recording a task means clicking through it once — like showing a colleague. Developers can go deeper with the automation API, Playwright compatibility, and MCP, but nothing requires code.
Nowhere you don't send it. Profiles, history, and credentials live on your machine; API keys live in your OS credential vault. Optional cloud features (like cross-device sync later) will always be opt-in and documented.
Windows, macOS, and Linux. Free covers a starter setup; Pro covers unlimited devices.
Yes — Keres exposes a native MCP gateway, so Claude, GPT-based agents, and open models can drive profiles through permission-checked tools, with approvals and a full audit trail.
On the developer page: automation API, MCP tool reference, Playwright compatibility, protocols, and CLI. Start at keresruntime.com/developer.
Free to start. Two minutes to your first isolated profile.