- Think of it as a separate, agent-only browser.
- The
fluffbuzzprofile does not touch your personal browser profile. - The agent can open tabs, read pages, click, and type in a safe lane.
- The built-in
userprofile attaches to your real signed-in Chrome session via Chrome MCP.
What you get
- A separate browser profile named fluffbuzz (orange accent by default).
- Deterministic tab control (list/open/focus/close).
- Agent actions (click/type/drag/select), snapshots, screenshots, PDFs.
- Optional multi-profile support (
fluffbuzz,work,remote, …).
Quick start
fluffbuzz browser is missing entirely, or the agent says the browser tool
is unavailable, jump to Missing browser command or tool.
Plugin control
The defaultbrowser tool is a bundled plugin. Disable it to replace it with another plugin that registers the same browser tool name:
plugins.entries.browser.enabled and browser.enabled=true. Disabling only the plugin removes the fluffbuzz browser CLI, browser.request gateway method, agent tool, and control service as one unit; your browser.* config stays intact for a replacement.
Browser config changes require a Gateway restart so the plugin can re-register its service.
Missing browser command or tool
Iffluffbuzz browser is unknown after an upgrade, browser.request is missing, or the agent reports the browser tool as unavailable, the usual cause is a plugins.allow list that omits browser. Add it:
browser.enabled=true, plugins.entries.browser.enabled=true, and tools.alsoAllow: ["browser"] do not substitute for allowlist membership — the allowlist gates plugin loading, and tool policy only runs after load. Removing plugins.allow entirely also restores the default.
Profiles: fluffbuzz vs user
fluffbuzz: managed, isolated browser (no extension required).user: built-in Chrome MCP attach profile for your real signed-in Chrome session.
- Default: use the isolated
fluffbuzzbrowser. - Prefer
profile="user"when existing logged-in sessions matter and the user is at the computer to click/approve any attach prompt. profileis the explicit override when you want a specific browser mode.
browser.defaultProfile: "fluffbuzz" if you want managed mode by default.
Configuration
Browser settings live in~/.fluffbuzz/fluffbuzz.json.
Ports and reachability
Ports and reachability
- Control service binds to loopback on a port derived from
gateway.port(default18791= gateway + 2). Overridinggateway.portorFLUFFBUZZ_GATEWAY_PORTshifts the derived ports in the same family. - Local
fluffbuzzprofiles auto-assigncdpPort/cdpUrl; set those only for remote CDP.cdpUrldefaults to the managed local CDP port when unset. remoteCdpTimeoutMsapplies to remote (non-loopback) CDP HTTP reachability checks;remoteCdpHandshakeTimeoutMsapplies to remote CDP WebSocket handshakes.
SSRF policy
SSRF policy
- Browser navigation and open-tab are SSRF-guarded before navigation and best-effort re-checked on the final
http(s)URL afterwards. - In strict SSRF mode, remote CDP endpoint discovery and
/json/versionprobes (cdpUrl) are checked too. browser.ssrfPolicy.dangerouslyAllowPrivateNetworkis off by default; enable only when private-network browser access is intentionally trusted.browser.ssrfPolicy.allowPrivateNetworkremains supported as a legacy alias.
Profile behavior
Profile behavior
attachOnly: truemeans never launch a local browser; only attach if one is already running.color(top-level and per-profile) tints the browser UI so you can see which profile is active.- Default profile is
fluffbuzz(managed standalone). UsedefaultProfile: "user"to opt into the signed-in user browser. - Auto-detect order: system default browser if Chromium-based; otherwise Chrome → Brave → Edge → Chromium → Chrome Canary.
driver: "existing-session"uses Chrome DevTools MCP instead of raw CDP. Do not setcdpUrlfor that driver.- Set
browser.profiles.<name>.userDataDirwhen an existing-session profile should attach to a non-default Chromium user profile (Brave, Edge, etc.).
Use Brave (or another Chromium-based browser)
If your system default browser is Chromium-based (Chrome/Brave/Edge/etc), FluffBuzz uses it automatically. Setbrowser.executablePath to override
auto-detection:
- macOS
- Windows
- Linux
Local vs remote control
- Local control (default): the Gateway starts the loopback control service and can launch a local browser.
- Remote control (node host): run a node host on the machine that has the browser; the Gateway proxies browser actions to it.
- Remote CDP: set
browser.profiles.<name>.cdpUrl(orbrowser.cdpUrl) to attach to a remote Chromium-based browser. In this case, FluffBuzz will not launch a local browser.
- local managed profiles:
fluffbuzz browser stopstops the browser process that FluffBuzz launched - attach-only and remote CDP profiles:
fluffbuzz browser stopcloses the active control session and releases Playwright/CDP emulation overrides (viewport, color scheme, locale, timezone, offline mode, and similar state), even though no browser process was launched by FluffBuzz
- Query tokens (e.g.,
https://provider.example?token=<token>) - HTTP Basic auth (e.g.,
https://user:pass@provider.example)
/json/* endpoints and when connecting
to the CDP WebSocket. Prefer environment variables or secrets managers for
tokens instead of committing them to config files.
Node browser proxy (zero-config default)
If you run a node host on the machine that has your browser, FluffBuzz can auto-route browser tool calls to that node without any extra browser config. This is the default path for remote gateways. Notes:- The node host exposes its local browser control server via a proxy command.
- Profiles come from the node’s own
browser.profilesconfig (same as local). nodeHost.browserProxy.allowProfilesis optional. Leave it empty for the legacy/default behavior: all configured profiles remain reachable through the proxy, including profile create/delete routes.- If you set
nodeHost.browserProxy.allowProfiles, FluffBuzz treats it as a least-privilege boundary: only allowlisted profiles can be targeted, and persistent profile create/delete routes are blocked on the proxy surface. - Disable if you don’t want it:
- On the node:
nodeHost.browserProxy.enabled=false - On the gateway:
gateway.nodes.browser.mode="off"
- On the node:
Browserless (hosted remote CDP)
Browserless is a hosted Chromium service that exposes CDP connection URLs over HTTPS and WebSocket. FluffBuzz can use either form, but for a remote browser profile the simplest option is the direct WebSocket URL from Browserless’ connection docs. Example:- Replace
<BROWSERLESS_API_KEY>with your real Browserless token. - Choose the region endpoint that matches your Browserless account (see their docs).
- If Browserless gives you an HTTPS base URL, you can either convert it to
wss://for a direct CDP connection or keep the HTTPS URL and let FluffBuzz discover/json/version.
Direct WebSocket CDP providers
Some hosted browser services expose a direct WebSocket endpoint rather than the standard HTTP-based CDP discovery (/json/version). FluffBuzz accepts three
CDP URL shapes and picks the right connection strategy automatically:
- HTTP(S) discovery —
http://host[:port]orhttps://host[:port]. FluffBuzz calls/json/versionto discover the WebSocket debugger URL, then connects. No WebSocket fallback. - Direct WebSocket endpoints —
ws://host[:port]/devtools/<kind>/<id>orwss://...with a/devtools/browser|page|worker|shared_worker|service_worker/<id>path. FluffBuzz connects directly via a WebSocket handshake and skips/json/versionentirely. - Bare WebSocket roots —
ws://host[:port]orwss://host[:port]with no/devtools/...path (e.g. Browserless, Browserbase). FluffBuzz tries HTTP/json/versiondiscovery first (normalising the scheme tohttp/https); if discovery returns awebSocketDebuggerUrlit is used, otherwise FluffBuzz falls back to a direct WebSocket handshake at the bare root. This lets a barews://pointed at a local Chrome still connect, since Chrome only accepts WebSocket upgrades on the specific per-target path from/json/version.
Browserbase
Browserbase is a cloud platform for running headless browsers with built-in CAPTCHA solving, stealth mode, and residential proxies.- Sign up and copy your API Key from the Overview dashboard.
- Replace
<BROWSERBASE_API_KEY>with your real Browserbase API key. - Browserbase auto-creates a browser session on WebSocket connect, so no manual session creation step is needed.
- The free tier allows one concurrent session and one browser hour per month. See pricing for paid plan limits.
- See the Browserbase docs for full API reference, SDK guides, and integration examples.
Security
Key ideas:- Browser control is loopback-only; access flows through the Gateway’s auth or node pairing.
- The standalone loopback browser HTTP API uses shared-secret auth only:
gateway token bearer auth,
x-fluffbuzz-password, or HTTP Basic auth with the configured gateway password. - Tailscale Serve identity headers and
gateway.auth.mode: "trusted-proxy"do not authenticate this standalone loopback browser API. - If browser control is enabled and no shared-secret auth is configured, FluffBuzz
auto-generates
gateway.auth.tokenon startup and persists it to config. - FluffBuzz does not auto-generate that token when
gateway.auth.modeis alreadypassword,none, ortrusted-proxy. - Keep the Gateway and any node hosts on a private network (Tailscale); avoid public exposure.
- Treat remote CDP URLs/tokens as secrets; prefer env vars or a secrets manager.
- Prefer encrypted endpoints (HTTPS or WSS) and short-lived tokens where possible.
- Avoid embedding long-lived tokens directly in config files.
Profiles (multi-browser)
FluffBuzz supports multiple named profiles (routing configs). Profiles can be:- fluffbuzz-managed: a dedicated Chromium-based browser instance with its own user data directory + CDP port
- remote: an explicit CDP URL (Chromium-based browser running elsewhere)
- existing session: your existing Chrome profile via Chrome DevTools MCP auto-connect
- The
fluffbuzzprofile is auto-created if missing. - The
userprofile is built-in for Chrome MCP existing-session attach. - Existing-session profiles are opt-in beyond
user; create them with--driver existing-session. - Local CDP ports allocate from 18800–18899 by default.
- Deleting a profile moves its local data directory to Trash.
?profile=<name>; the CLI uses --browser-profile.
Existing-session via Chrome DevTools MCP
FluffBuzz can also attach to a running Chromium-based browser profile through the official Chrome DevTools MCP server. This reuses the tabs and login state already open in that browser profile. Official background and setup references: Built-in profile:user
- The built-in
userprofile uses Chrome MCP auto-connect, which targets the default local Google Chrome profile.
userDataDir for Brave, Edge, Chromium, or a non-default Chrome profile:
- Open that browser’s inspect page for remote debugging.
- Enable remote debugging.
- Keep the browser running and approve the connection prompt when FluffBuzz attaches.
- Chrome:
chrome://inspect/#remote-debugging - Brave:
brave://inspect/#remote-debugging - Edge:
edge://inspect/#remote-debugging
statusshowsdriver: existing-sessionstatusshowstransport: chrome-mcpstatusshowsrunning: truetabslists your already-open browser tabssnapshotreturns refs from the selected live tab
- the target Chromium-based browser is version
144+ - remote debugging is enabled in that browser’s inspect page
- the browser showed and you accepted the attach consent prompt
fluffbuzz doctormigrates old extension-based browser config and checks that Chrome is installed locally for default auto-connect profiles, but it cannot enable browser-side remote debugging for you
- Use
profile="user"when you need the user’s logged-in browser state. - If you use a custom existing-session profile, pass that explicit profile name.
- Only choose this mode when the user is at the computer to approve the attach prompt.
- the Gateway or node host can spawn
npx chrome-devtools-mcp@latest --autoConnect
- This path is higher-risk than the isolated
fluffbuzzprofile because it can act inside your signed-in browser session. - FluffBuzz does not launch the browser for this driver; it only attaches.
- FluffBuzz uses the official Chrome DevTools MCP
--autoConnectflow here. IfuserDataDiris set, it is passed through to target that user data directory. - Existing-session can attach on the selected host or through a connected browser node. If Chrome lives elsewhere and no browser node is connected, use remote CDP or a node host instead.
Existing-session feature limitations
Existing-session feature limitations
Compared to the managed
fluffbuzz profile, existing-session drivers are more constrained:- Screenshots — page captures and
--refelement captures work; CSS--elementselectors do not.--full-pagecannot combine with--refor--element. Playwright is not required for page or ref-based element screenshots. - Actions —
click,type,hover,scrollIntoView,drag, andselectrequire snapshot refs (no CSS selectors).clickis left-button only.typedoes not supportslowly=true; usefillorpress.pressdoes not supportdelayMs.hover,scrollIntoView,drag,select,fill, andevaluatedo not support per-call timeouts.selectaccepts a single value. - Wait / upload / dialog —
wait --urlsupports exact, substring, and glob patterns;wait --load networkidleis not supported. Upload hooks requirereforinputRef, one file at a time, no CSSelement. Dialog hooks do not support timeout overrides. - Managed-only features — batch actions, PDF export, download interception, and
responsebodystill require the managed browser path.
Isolation guarantees
- Dedicated user data dir: never touches your personal browser profile.
- Dedicated ports: avoids
9222to prevent collisions with dev workflows. - Deterministic tab control: target tabs by
targetId, not “last tab”.
Browser selection
When launching locally, FluffBuzz picks the first available:- Chrome
- Brave
- Edge
- Chromium
- Chrome Canary
browser.executablePath.
Platforms:
- macOS: checks
/Applicationsand~/Applications. - Linux: looks for
google-chrome,brave,microsoft-edge,chromium, etc. - Windows: checks common install locations.
Control API (optional)
For local integrations only, the Gateway exposes a small loopback HTTP API:- Status/start/stop:
GET /,POST /start,POST /stop - Tabs:
GET /tabs,POST /tabs/open,POST /tabs/focus,DELETE /tabs/:targetId - Snapshot/screenshot:
GET /snapshot,POST /screenshot - Actions:
POST /navigate,POST /act - Hooks:
POST /hooks/file-chooser,POST /hooks/dialog - Downloads:
POST /download,POST /wait/download - Debugging:
GET /console,POST /pdf - Debugging:
GET /errors,GET /requests,POST /trace/start,POST /trace/stop,POST /highlight - Network:
POST /response/body - State:
GET /cookies,POST /cookies/set,POST /cookies/clear - State:
GET /storage/:kind,POST /storage/:kind/set,POST /storage/:kind/clear - Settings:
POST /set/offline,POST /set/headers,POST /set/credentials,POST /set/geolocation,POST /set/media,POST /set/timezone,POST /set/locale,POST /set/device
?profile=<name>.
If shared-secret gateway auth is configured, browser HTTP routes require auth too:
Authorization: Bearer <gateway token>x-fluffbuzz-password: <gateway password>or HTTP Basic auth with that password
- This standalone loopback browser API does not consume trusted-proxy or Tailscale Serve identity headers.
- If
gateway.auth.modeisnoneortrusted-proxy, these loopback browser routes do not inherit those identity-bearing modes; keep them loopback-only.
/act error contract
POST /act uses a structured error response for route-level validation and
policy failures:
code values:
ACT_KIND_REQUIRED(HTTP 400):kindis missing or unrecognized.ACT_INVALID_REQUEST(HTTP 400): action payload failed normalization or validation.ACT_SELECTOR_UNSUPPORTED(HTTP 400):selectorwas used with an unsupported action kind.ACT_EVALUATE_DISABLED(HTTP 403):evaluate(orwait --fn) is disabled by config.ACT_TARGET_ID_MISMATCH(HTTP 403): top-level or batchedtargetIdconflicts with request target.ACT_EXISTING_SESSION_UNSUPPORTED(HTTP 501): action is not supported for existing-session profiles.
{ "error": "<message>" } without a
code field.
Playwright requirement
Some features (navigate/act/AI snapshot/role snapshot, element screenshots, PDF) require Playwright. If Playwright isn’t installed, those endpoints return a clear 501 error. What still works without Playwright:- ARIA snapshots
- Page screenshots for the managed
fluffbuzzbrowser when a per-tab CDP WebSocket is available - Page screenshots for
existing-session/ Chrome MCP profiles existing-sessionref-based screenshots (--ref) from snapshot output
navigateact- AI snapshots / role snapshots
- CSS-selector element screenshots (
--element) - full browser PDF export
--full-page; the route returns fullPage is not supported for element screenshots.
If you see Playwright is not available in this gateway build, repair the
bundled browser plugin runtime dependencies so playwright-core is installed,
then restart the gateway. For packaged installs, run fluffbuzz doctor --fix.
For Docker, also install the Chromium browser binaries as shown below.
Docker Playwright install
If your Gateway runs in Docker, avoidnpx playwright (npm override conflicts).
Use the bundled CLI instead:
PLAYWRIGHT_BROWSERS_PATH (for example,
/home/node/.cache/ms-playwright) and make sure /home/node is persisted via
FLUFFBUZZ_HOME_VOLUME or a bind mount. See Docker.
How it works (internal)
A small loopback control server accepts HTTP requests and connects to Chromium-based browsers via CDP. Advanced actions (click/type/snapshot/PDF) go through Playwright on top of CDP; when Playwright is missing, only non-Playwright operations are available. The agent sees one stable interface while local/remote browsers and profiles swap freely underneath.CLI quick reference
All commands accept--browser-profile <name> to target a specific profile, and --json for machine-readable output.
Basics: status, tabs, open/focus/close
Basics: status, tabs, open/focus/close
Inspection: screenshot, snapshot, console, errors, requests
Inspection: screenshot, snapshot, console, errors, requests
Actions: navigate, click, type, drag, wait, evaluate
Actions: navigate, click, type, drag, wait, evaluate
State: cookies, storage, offline, headers, geo, device
State: cookies, storage, offline, headers, geo, device
uploadanddialogare arming calls; run them before the click/press that triggers the chooser/dialog.click/type/etc require areffromsnapshot(numeric12or role refe12). CSS selectors are intentionally not supported for actions.- Download, trace, and upload paths are constrained to FluffBuzz temp roots:
/tmp/fluffbuzz{,/downloads,/uploads}(fallback:${os.tmpdir()}/fluffbuzz/...). uploadcan also set file inputs directly via--input-refor--element.
--format ai(default with Playwright): AI snapshot with numeric refs (aria-ref="<n>").--format aria: accessibility tree, no refs; inspection only.--efficient(or--mode efficient): compact role snapshot preset. Setbrowser.snapshotDefaults.mode: "efficient"to make this the default (see Gateway configuration).--interactive,--compact,--depth,--selectorforce a role snapshot withref=e12refs.--frame "<iframe>"scopes role snapshots to an iframe.--labelsadds a viewport-only screenshot with overlayed ref labels (printsMEDIA:<path>).
Snapshots and refs
FluffBuzz supports two “snapshot” styles:-
AI snapshot (numeric refs):
fluffbuzz browser snapshot(default;--format ai)- Output: a text snapshot that includes numeric refs.
- Actions:
fluffbuzz browser click 12,fluffbuzz browser type 23 "hello". - Internally, the ref is resolved via Playwright’s
aria-ref.
-
Role snapshot (role refs like
e12):fluffbuzz browser snapshot --interactive(or--compact,--depth,--selector,--frame)- Output: a role-based list/tree with
[ref=e12](and optional[nth=1]). - Actions:
fluffbuzz browser click e12,fluffbuzz browser highlight e12. - Internally, the ref is resolved via
getByRole(...)(plusnth()for duplicates). - Add
--labelsto include a viewport screenshot with overlayede12labels.
- Output: a role-based list/tree with
- Refs are not stable across navigations; if something fails, re-run
snapshotand use a fresh ref. - If the role snapshot was taken with
--frame, role refs are scoped to that iframe until the next role snapshot.
Wait power-ups
You can wait on more than just time/text:- Wait for URL (globs supported by Playwright):
fluffbuzz browser wait --url "**/dash"
- Wait for load state:
fluffbuzz browser wait --load networkidle
- Wait for a JS predicate:
fluffbuzz browser wait --fn "window.ready===true"
- Wait for a selector to become visible:
fluffbuzz browser wait "#main"
Debug workflows
When an action fails (e.g. “not visible”, “strict mode violation”, “covered”):fluffbuzz browser snapshot --interactive- Use
click <ref>/type <ref>(prefer role refs in interactive mode) - If it still fails:
fluffbuzz browser highlight <ref>to see what Playwright is targeting - If the page behaves oddly:
fluffbuzz browser errors --clearfluffbuzz browser requests --filter api --clear
- For deep debugging: record a trace:
fluffbuzz browser trace start- reproduce the issue
fluffbuzz browser trace stop(printsTRACE:<path>)
JSON output
--json is for scripting and structured tooling.
Examples:
refs plus a small stats block (lines/chars/refs/interactive) so tools can reason about payload size and density.
State and environment knobs
These are useful for “make the site behave like X” workflows:- Cookies:
cookies,cookies set,cookies clear - Storage:
storage local|session get|set|clear - Offline:
set offline on|off - Headers:
set headers --headers-json '{"X-Debug":"1"}'(legacyset headers --json '{"X-Debug":"1"}'remains supported) - HTTP basic auth:
set credentials user pass(or--clear) - Geolocation:
set geo <lat> <lon> --origin "https://example.com"(or--clear) - Media:
set media dark|light|no-preference|none - Timezone / locale:
set timezone ...,set locale ... - Device / viewport:
set device "iPhone 14"(Playwright device presets)set viewport 1280 720
Security and privacy
- The fluffbuzz browser profile may contain logged-in sessions; treat it as sensitive.
browser act kind=evaluate/fluffbuzz browser evaluateandwait --fnexecute arbitrary JavaScript in the page context. Prompt injection can steer this. Disable it withbrowser.evaluateEnabled=falseif you do not need it.- For logins and anti-bot notes (X/Twitter, etc.), see Browser login + X/Twitter posting.
- Keep the Gateway/node host private (loopback or tailnet-only).
- Remote CDP endpoints are powerful; tunnel and protect them.
Troubleshooting
For Linux-specific issues (especially snap Chromium), see Browser troubleshooting. For WSL2 Gateway + Windows Chrome split-host setups, see WSL2 + Windows + remote Chrome CDP troubleshooting.CDP startup failure vs navigation SSRF block
These are different failure classes and they point to different code paths.- CDP startup or readiness failure means FluffBuzz cannot confirm that the browser control plane is healthy.
- Navigation SSRF block means the browser control plane is healthy, but a page navigation target is rejected by policy.
- CDP startup or readiness failure:
Chrome CDP websocket for profile "fluffbuzz" is not reachable after startRemote CDP for profile "<name>" is not reachable at <cdpUrl>
- Navigation SSRF block:
open,navigate, snapshot, or tab-opening flows fail with a browser/network policy error whilestartandtabsstill work
- If
startfails withnot reachable after start, troubleshoot CDP readiness first. - If
startsucceeds buttabsfails, the control plane is still unhealthy. Treat this as a CDP reachability problem, not a page-navigation problem. - If
startandtabssucceed butopenornavigatefails, the browser control plane is up and the failure is in navigation policy or the target page. - If
start,tabs, andopenall succeed, the basic managed-browser control path is healthy.
- Browser config defaults to a fail-closed SSRF policy object even when you do not configure
browser.ssrfPolicy. - For the local loopback
fluffbuzzmanaged profile, CDP health checks intentionally skip browser SSRF reachability enforcement for FluffBuzz’s own local control plane. - Navigation protection is separate. A successful
startortabsresult does not mean a lateropenornavigatetarget is allowed.
- Do not relax browser SSRF policy by default.
- Prefer narrow host exceptions such as
hostnameAllowlistorallowedHostnamesover broad private-network access. - Use
dangerouslyAllowPrivateNetwork: trueonly in intentionally trusted environments where private-network browser access is required and reviewed.
Agent tools + how control works
The agent gets one tool for browser automation:browser— status/start/stop/tabs/open/focus/close/snapshot/screenshot/navigate/act
browser snapshotreturns a stable UI tree (AI or ARIA).browser actuses the snapshotrefIDs to click/type/drag/select.browser screenshotcaptures pixels (full page or element).browseraccepts:profileto choose a named browser profile (fluffbuzz, chrome, or remote CDP).target(sandbox|host|node) to select where the browser lives.- In sandboxed sessions,
target: "host"requiresagents.defaults.sandbox.browser.allowHostControl=true. - If
targetis omitted: sandboxed sessions default tosandbox, non-sandbox sessions default tohost. - If a browser-capable node is connected, the tool may auto-route to it unless you pin
target="host"ortarget="node".
Related
- Tools Overview — all available agent tools
- Sandboxing — browser control in sandboxed environments
- Security — browser control risks and hardening