-Botloft keeps a crew of [Claude Code](https://code.claude.com) bots running on your computer. Each bot has
-its own folder, memory and conversation. They pass work to each other, keep going after you close the
+Botloft keeps a crew of bots running on your computer, each on the AI coding agent you choose:
+[Claude Code](https://code.claude.com), or [Antigravity](https://antigravity.google) as an experimental
+option, with Codex next. Each bot has its own folder, memory and conversation. They pass work to each other, keep going after you close the
window and pick up where they left off after a restart. You talk to each one like in a chat app, and
you decide what they may do.
@@ -55,8 +56,8 @@ you decide what they may do.
- **Leave them running.** Bots keep working after you close Botloft. They start when you sign in,
come back soon after a crash and keep the computer awake while they work. Botloft
waits near the clock and tells you when a bot needs you.
-- **Choose per bot.** How much it asks before acting, and which Claude model it uses. See how much
- of your plan's usage is left.
+- **Choose per bot.** Which agent it runs on, which model it uses, and how much it asks before
+ acting. With Claude Code, see how much of your plan's usage is left.
- **Back up and move.** Save a sealed backup file, or sign in with your e-mail (no password) to keep
a light copy in the cloud and restore it on another computer. The server cannot read it. Turn on
automatic copies, every day or week and only when something changed, if you want them to go by
@@ -76,10 +77,26 @@ you decide what they may do.
Botloft speaks English, Portuguese (Brazil) and Spanish, follows your light or dark theme and updates
itself.
+## Agents
+
+| | Claude Code | Antigravity (experimental) |
+|---|---|---|
+| Chat, crew messages, tasks, routines, the live browser | yes | chat, crew messages and tasks; the rest is not checked yet |
+| Asks you before acting | yes: each tool, with "always allow" | no: it cannot ask. It works with files in its own folders; commands you list per bot (or "full access") |
+| Images | yes | yes, by file |
+| Model | your plan's default, or one you pick | its own list; the effort is part of the name |
+| Usage meter and "Compact" | yes | an estimate of the conversation size; no compaction |
+| Isolation between crews | yes | folders are fenced for its file tools, not for commands you allow |
+
+Claude Code is the full experience. Antigravity is off until you add `experimental_agents = ["agy"]`
+to Botloft's `config.toml`; then a new bot can be made on it, and Settings has "Agent for new
+bots". Codex is the next agent. See [spec section 30](docs/spec.md) for what was checked.
+
## Get started
-You need [Claude Code](https://code.claude.com/docs/en/setup), installed and signed in once with
-your Claude account. Bots run on your own Claude plan.
+You need at least one agent installed and signed in once, with your own account:
+[Claude Code](https://code.claude.com/docs/en/setup), or Antigravity (experimental, see
+[Agents](#agents)). Bots run on your own plan with that agent.
1. Download Botloft for your computer from the
[latest release](https://github.com/httpsphl/botloft/releases/latest):
@@ -112,15 +129,16 @@ app. Your bots and their files stay either way (see
flowchart LR
app["Botloft appTauri + React"] -- "JSON-RPC over WebSocket
127.0.0.1 only" --> daemon["botloftd
Rust daemon"] daemon --> db[("SQLite")] - daemon -- "stdin and stdout
stream-json" --> bots["One claude -p
per bot"] + daemon -- "stdin and stdout
stream-json" --> bots["One agent process
per bot"] bots -- "Botloft's MCP tools
messages, tasks, requests" --> daemon ``` - **The daemon is the source of truth.** Closing the app never stops a bot. `botloftd` runs as a scheduled task of your Windows user, a systemd user service on Linux or a launch agent on macOS: it starts when you sign in and comes back soon if it dies. -- **Claude Code is the runtime.** Each bot is a real `claude -p` session with its own workspace, - memory (`CLAUDE.md`) and conversation. There is no custom agent SDK and no API key to manage. +- **Your agent is the runtime.** Each bot is a real session of the agent you chose (`claude -p` for + Claude Code, `agy -p` for Antigravity) with its own workspace, memory and conversation. There is + no custom agent SDK and no API key to manage. - **One conversation per bot.** Your messages, messages from other bots and notices from the daemon reach the bot in order. What it does comes back as a chat: replies, tool use and requests you allow or deny. @@ -134,7 +152,7 @@ The full design, from the protocol to the states of a bot, lives in [`docs/spec. - **Local first.** The daemon listens on `127.0.0.1`, and there is no telemetry. An account is optional: without one, Botloft uses the network for nothing beyond updates and what bots send to - Claude, which goes through Claude Code as when you use it yourself. With one, the server only ever + the agent's provider, which goes through the agent as when you use it yourself. With one, the server only ever holds a backup sealed with a password that never leaves your computer (see the FAQ). - **You decide what bots may do.** Per bot: ask before anything, accept edits, plan only, or decide on its own what needs your OK. Unless it decides on its own, a bot also asks before its first visit @@ -143,7 +161,7 @@ The full design, from the protocol to the states of a bot, lives in [`docs/spec. Each bot gets its own token per start, and the daemon keeps only its SHA-256 hash. - **Botloft is not a sandbox.** Every bot runs as your user account, so isolation between bots is cooperative: a bot can read other bots' folders if it decides to. Give bots only the permissions - you would give Claude Code directly. + you would give that agent directly. What Botloft sends over the network, what it changes on your computer and how to remove it are in the [code signing policy](docs/code-signing-policy.md). @@ -154,8 +172,8 @@ the [code signing policy](docs/code-signing-policy.md).
@@ -212,8 +230,8 @@ Windows came first and has had the most use; if something does not work elsewher
## Status
@@ -228,8 +246,8 @@ and a code-signed installer. `main` can be ahead of the latest release.
Requirements: Rust stable (the MSVC toolchain on Windows), Node.js 24 and pnpm. On Linux, the app
also needs the desktop libraries Tauri uses (on Ubuntu: `libwebkit2gtk-4.1-dev`,
-`libayatana-appindicator3-dev`, `librsvg2-dev`, `libxdo-dev`). Running real bots also needs Claude
-Code (the native `claude` executable) installed and signed in.
+`libayatana-appindicator3-dev`, `librsvg2-dev`, `libxdo-dev`). Running real bots also needs an agent
+installed and signed in: Claude Code (the native `claude` executable) or Antigravity (`agy`).
```powershell
# Daemon and libraries
diff --git a/app/src/features/account/GeneralSettings.tsx b/app/src/features/account/GeneralSettings.tsx
index 874f52ee..054c988b 100644
--- a/app/src/features/account/GeneralSettings.tsx
+++ b/app/src/features/account/GeneralSettings.tsx
@@ -9,12 +9,16 @@ import {
useLocale,
useT,
} from "../../i18n";
+import type { AgentKind } from "../../lib/protocol.gen";
import { SYSTEM } from "../../lib/system";
import { prefs, usePref } from "../../shell/prefs";
+import { useApp } from "../../store/context";
import { Select } from "../../ui/Select";
import { Section, Toggle } from "./settingsParts";
import { useDaemonSettings } from "./useDaemonSettings";
+const NO_AGENTS: AgentKind[] = [];
+
export function GeneralSettings() {
const t = useT();
const s = t.account.settings;
@@ -23,6 +27,7 @@ export function GeneralSettings() {
const openAtSignIn = usePref(prefs.openAtSignIn);
const nearClock = whenClosed === "keep" && tray;
const { settings, change } = useDaemonSettings();
+ const enabledAgents = useApp((state) => state.system?.enabledAgents ?? NO_AGENTS);
const startWithWindows = settings?.startWithWindows ?? true;
const { choice: locale } = useLocale();
const systemName = LOCALES.find((entry) => entry.id === systemLocale())?.name ?? "English";
@@ -36,6 +41,20 @@ export function GeneralSettings() {
+ {settings && enabledAgents.length > 1 && (
+
+
+ )}
{
startWithWindows: false,
keepAwake: false,
approvalWaitMinutes: 60,
+ defaultAgent: "claude",
}),
);
expect(
diff --git a/app/src/features/account/settingsIndex.ts b/app/src/features/account/settingsIndex.ts
index 454d1591..e64b9194 100644
--- a/app/src/features/account/settingsIndex.ts
+++ b/app/src/features/account/settingsIndex.ts
@@ -32,6 +32,7 @@ export function settingsIndex(t: ReturnType): Entry[] {
...entry(
"general",
s.language,
+ s.defaultAgent,
s.background,
s.keepWorking,
s.tray,
diff --git a/app/src/features/account/useDaemonSettings.ts b/app/src/features/account/useDaemonSettings.ts
index abd0c489..08854d63 100644
--- a/app/src/features/account/useDaemonSettings.ts
+++ b/app/src/features/account/useDaemonSettings.ts
@@ -23,6 +23,7 @@ export function useDaemonSettings(): {
startWithWindows: update.startWithWindows ?? before.startWithWindows,
keepAwake: update.keepAwake ?? before.keepAwake,
approvalWaitMinutes: update.approvalWaitMinutes ?? before.approvalWaitMinutes,
+ defaultAgent: update.defaultAgent ?? before.defaultAgent,
});
}
attempt(failed, async () => putSettings(await api.call("settings.update", update))).then(
diff --git a/app/src/features/bots/BotDialog.tsx b/app/src/features/bots/BotDialog.tsx
index b94e481a..0bc8646d 100644
--- a/app/src/features/bots/BotDialog.tsx
+++ b/app/src/features/bots/BotDialog.tsx
@@ -14,6 +14,7 @@ import { useApi, useApp } from "../../store/context";
import { Button } from "../../ui/Button";
import { Dialog } from "../../ui/Dialog";
import { SelectField, TextArea, TextField } from "../../ui/Field";
+import { Toggle } from "../account/settingsParts";
import { BotAvatar } from "./BotAvatar";
import { ColorPicker } from "./ColorPicker";
@@ -41,8 +42,13 @@ export function BotDialog(props: Props) {
const [commands, setCommands] = useState((editing?.allowedCommands ?? []).join("\n"));
// Which agent runs the bot is chosen once, when it is made (spec 30).
const agents = useApp((state) => state.system?.enabledAgents ?? CLAUDE_ONLY);
- const [agent, setAgent] = useState("claude");
+ const preferred = useApp((state) => state.settings?.defaultAgent ?? "claude");
+ const [agent, setAgent] = useState(agents.includes(preferred) ? preferred : "claude");
const runsOn: AgentKind = editing?.agent ?? agent;
+ const fullAccess = commands
+ .split("\n")
+ .map((line) => line.trim())
+ .includes("*");
const [error, setError] = useState(null);
const [busy, setBusy] = useState(false);
@@ -148,14 +154,24 @@ export function BotDialog(props: Props) {
)}
{editing && editing.agent !== "claude" && (
-