Spawn a detached `git pull --ff-only` in the background after each
site command. Non-blocking, no output, keeps community adapters fresh
without user intervention.
Closes#20
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Only show domains with >= 5 visits
- Exclude localhost and non-domain strings (e.g. Chrome extension IDs)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The previous patch only modified extension/manifest.json (build output,
gitignored) instead of packages/extension/manifest.json (source). Every
vite build overwrote the fix. Now the source manifest has the history
permission so it survives rebuilds.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Quick Start: replace `site list` with `site recommend`
- Update platform count: 36 platforms, 103 commands
- Add Xueqiu to Finance category
- Document --jq and site info with examples
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
New commands:
- `history search/domains` — query Chrome browsing history via chrome.history API
- `site recommend` — cross-reference history with adapters for personalized discovery
- `site info <name>` — expose full adapter @meta (args, example, domain)
- `--jq <expr>` — inline jq-like JSON filtering (inspired by gh CLI)
Agent UX improvements:
- --help grouped by intent, "开始使用" section first with site examples
- --help top hint explicitly guides Agents to use site commands over browser automation
- `open` command shows 💡 hint when the URL has available site adapters
- `--jq` auto-strips `.data.` prefix for Agent tolerance
- `site update` post-install nudge for site recommend
- AGENTS.md with full UX writing spec for contributors
Closes#21, closes#23, relates #24
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Document the --host flag for daemon bind address customization:
- 127.0.0.1 for IPv4-only (fixes macOS IPv6 issues, ref #18)
- 0.0.0.0 for remote access via Tailscale / ZeroTier
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Update numbers: 10 platforms → 35, 50+ → 97 commands
- Add "The idea" section: browser IS the API, not the other way around
- Add "What this means for AI agents": files + terminal + entire internet
- Add "10 minutes to add any website" with tier system and parallel AI proof
- Full platform table by category
- Bilingual (EN + ZH)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Group commands by function (site adapters, navigation, interaction,
page info, network/debug, daemon) instead of a flat list.
Move site adapter section to the top with usage examples, making
the most powerful feature immediately visible.
Closes#13
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Lead with 50+ site commands across 10 platforms instead of
generic browser automation. That's what makes bb-browser different.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
MCP server now sends instructions on connect so agents immediately
understand what bb-browser can do (capabilities, site adapters, etc).
New top-level `bb-browser guide` command teaches how to turn any
website into a site adapter — reverse engineer, write, test, contribute.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
reportHint now shows two options for agents:
1. gh issue create --repo epiral/bb-sites
2. bb-browser site github/issue-create epiral/bb-sites
Agents choose whichever tool is available.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When a site adapter fails, include a reportHint field telling the
agent how to file an issue via `gh issue create --repo epiral/bb-sites`.
Agents decide themselves whether to use gh CLI or bb-browser's own
github site adapter.
Remove site report subcommand — agents handle reporting autonomously.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Check both error and hint fields for auth keywords.
When detected, replace with actionable login URL hint.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When a site adapter returns 401/403/unauthorized errors, automatically
suggest the user to log in to the site's domain in their browser.
Works in both CLI and JSON output modes.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
MCP server now spawns daemon automatically on first tool call,
same as CLI does. Only shows setup hint for extension not connected.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Show actionable setup guide when:
- Daemon not running (ECONNREFUSED)
- Extension not connected (HTTP 503)
Hints include install steps and GitHub releases link, guiding both
users and AI agents to complete the setup.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
`bb-browser --mcp` spawns the MCP server with stdio transport.
Users can configure AI tools with: npx -y bb-browser --mcp
Also fixes root tsup bundling to properly exclude ws while
bundling all other deps (MCP SDK, ajv, zod).
Bump version to 0.4.1.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Bump version to 0.4.0
- Add MCP runtime deps (ajv, ajv-formats, @modelcontextprotocol/sdk)
- Bundle all deps in root tsup for npx compatibility
- Fix repo URL to epiral/bb-browser
- Add mcp keyword
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace broken `import("@bb-browser/daemon")` with `spawn(getDaemonPath())`
so `bb-browser daemon` works both in dev monorepo and npm install scenarios.
- Export getDaemonPath() from daemon-manager
- daemon command uses spawn with stdio: "inherit" for foreground mode
- Pass --host flag through to daemon process
Fixes#6, fixes#10
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- parseArgs: remove `&& !result.command` guard so --tab value is not
leaked into command args (e.g. network filter)
- site: respect --tab, stop calling tab_select (no foreground steal)
- fetch: stop calling tab_select (no foreground steal)
- site --json: wrap output in standard {id, success, data} envelope
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- README.md: English (default)
- README.zh-CN.md: Chinese
- Cross-linked via header
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Rename narrative: bb = BadBoy Browser
- Tagline: "Your browser is the API. No keys. No bots. No scrapers."
- Add fetch, site, network --with-body to command reference
- Bilingual: English (default) + Chinese section
- Reframe positioning: not automation, not scraping — being the browser
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Via CDP Network.loadingFinished + Network.getResponseBody, capture:
- Request headers and body (POST data)
- Response headers, body, and mimeType
- Memory management: 256KB/response, 64KB/request, 8MB/tab total
- Eviction: oldest bodies removed first when over budget
- Graceful error handling: bodyError field for failed captures
This enables a new paradigm for site adapters: instead of reverse-engineering
API signing (like Xiaohongshu's X-s), just navigate to a page and read the
response bodies from requests the page already made with proper authentication.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- `bb-browser fetch <url>`: browser-native fetch with auto same-origin tab routing.
Like curl but with browser login state. No API keys, no anti-bot.
- `bb-browser recipe`: manage and run community/private fetch recipes.
- recipe list: scan ~/.bb-browser/recipes/ (local) and ~/.bb-browser/bb-recipes/ (community)
- recipe search <query>: keyword search across recipes
- recipe run <name> [args]: execute a recipe via eval
- recipe update: git clone/pull community recipes from epiral/bb-recipes
- recipe <name> [args]: shorthand for recipe run
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
yan5xu