gawkbot

open source grok bot. gawkbots use local VMs as computers (cloud option available) and build microapps to manage all your work.

https://github.com/user-attachments/assets/84c4b8cf-fb5d-4e9b-b720-213648c2e20b

gawkbot lets anyone turn their manual workflows into microapps across 1200+ integrations in minutes. Describe the job in one sentence — or demo it once on a call — and your AI builds the bot that runs it: its own screen, its own schedule, its own tools, with a human approval gate on everything it sends. Runs local, on your machine, on your account.

gawkbots automate your menial work via AI models and build you microapps to manage the outcome, so that you have a false sense of control.

grok (verb) — to understand something profoundly and intuitively. gawk (verb) — to stare openly and stupidly.

It is named after the second one. Not after the bots. The bots are working. You are the one with the dashboard open.

The open source Grok Bot

gawkbot is an open source Grok Bot: always-on AI bots on your own machine instead of xAI's cloud, free, on the coding bot you already pay for, with an approval gate on every external action. The honest version, row by row:

Full comparison, where Grok Bot is better, and the other open source alternatives (Rakazo, OpenMausBot, OpenBot): gawk.bot/open-source-grok-bot. Website: gawk.bot.

Get Started

Prerequisites: one bot CLI, signed in — Claude Code by default, or Codex CLI / Opencode. The first-run screen verifies your runtime before anything else happens.

npx gawkbot

That's it. The browser opens, you verify your runtime, name your office, and hand off your first workflow — you land on your first bot being built, live.

Prefer a global install?

npm install -g gawkbot && gawkbot

Building from source (requires Go and Bun):

git clone https://github.com/najmuzzaman-mohammad/gawkbot.git
cd gawkbot
cd web && bun install && bun run build && cd ..
go build -o gawkbot ./cmd/wuphf
./gawkbot

Routine execution runs on a small sidecar service (agent/). The broker finds and supervises it automatically on source checkouts (set WUPHF_AGENT_DIR to point elsewhere, or WUPHF_AGENT_URL if you manage it yourself).

What you get

Every bot ships with all six. Not a chatbot in a trench coat.

If your workflow names a system that is not connected yet ("audit our HubSpot"), gawkbot asks before building — build against live workspace data now, or hold while you connect. It never silently re-scopes your job.

Setup prompt (for AI bots)

Paste this into Claude Code, Codex, or Cursor and let your bot drive the install:

Set up https://github.com/najmuzzaman-mohammad/gawkbot for me. Read `README.md`
first, then run `npx gawkbot` — the web UI opens at http://localhost:7891.

Walk the onboarding: verify the runtime, name the office, and start the first
workflow. Confirm you land on a bot being built (a live build feed beside a
chat), and that when it finishes the bot shows tabs for UI, Routines, Tools,
Data, Knowledge, and Integrations.

For agent conventions read `AGENTS.md`; for internals read `ARCHITECTURE.md`;
for forking read `FORKING.md`.

Options

Local models and custom endpoints

For custom OpenAI-compatible endpoints (LiteLLM, local proxies, Ollama):

WUPHF_OLLAMA_BASE_URL="http://127.0.0.1:20128/v1" \
WUPHF_OLLAMA_MODEL="openai/gpt-5.4-mini" \
gawkbot --provider ollama --no-open

--provider opencode shells out to the opencode CLI binary; MLX-LM and Ollama can be set up from the first-run screen with no cloud key at all.

Other runtimes

Already running Hermes Bot or an OpenClaw gateway? Point bots at them with --provider hermes-agent (default http://127.0.0.1:8642/v1) or --provider openclaw-http (default http://127.0.0.1:18789/v1). Endpoints, models, and auth are overridable via WUPHF_HERMES_AGENT_* / WUPHF_OPENCLAW_HTTP_* env vars or provider_endpoints in config.

Memory: the company brain

gawkbot ships with built-in memory — no backend choice, no API key. Your workspace state lives in local files you can cat: bot knowledge, run transcripts, and the company brain under ~/.wuphf/. Knowledge pages are synthesized with citations back to their sources, so you can check the receipts on anything a bot claims.

Other Commands

gawkbot init                    # First-time setup
gawkbot share                   # Invite one team member over Tailscale/WireGuard
gawkbot shred                   # Delete workspace state and reopen onboarding
gawkbot workspace list          # Run multiple isolated workspaces side by side
gawkbot workspace switch <name> # Flip the active workspace

Share With a Team Member

Two ways to invite a teammate, both from the CLI:

Private network — Tailscale or WireGuard. Both machines on the same mesh; the invite never leaves the network:

gawkbot share

Public tunnel — no shared network needed. The broker can spin up a Cloudflare quick tunnel (POST /api/share/tunnel/start; the trycloudflare URL is paired with a 6-digit passcode, invites are one-use and expire in 24 hours, and the join handler is rate-limited per source IP). cloudflared ships with the npm install (pinned SHA256 per platform). The one-click button for this is being resurfaced in the operator shell — until then the endpoint is the path.

For the full walkthrough, see Share gawkbot With a Team Member.

External Actions

Bots act through two providers — pick whichever fits:

• One CLI (default, local-first): actions execute through a local CLI on your machine; credentials never leave it.
• Composio (cloud-hosted OAuth): connect Gmail, Slack, HubSpot, and the rest of the 1200+ catalog from any bot's Integrations tab. Connections are shared across the office.

Either way, the approval gate holds every external write until you approve it.

Privacy & Telemetry

gawkbot can send anonymous product analytics and session recordings (with typed text masked) to help us improve it. This is optional, controlled by you, and off unless a PostHog key is configured — a stock source build and every fork ship with no key, so they never phone home.

Two independent toggles (onboarding and Settings), both on by default, both reversible at any time:

• Product analytics — anonymous usage events: which flows are used, where people get stuck, error counts. We send counts and shapes only, never your content (no message text, task titles, customer data, or secrets).
• Session recording — recordings mask everything you type (maskAllInputs: true): passwords, API keys, and any form field are obscured. We capture layout, clicks, and navigation to fix rough edges.

No autocapture, no cookies (localStorage only). Self-hosted operators can point at their own PostHog (WUPHF_POSTHOG_KEY / WUPHF_POSTHOG_HOST) or leave the key unset to keep gawkbot fully dormant. Full taxonomy and policy: docs/specs/product-analytics.md.

Why gawkbot

Claim Status

Every claim in this README, grounded to the code that makes it true.

Legend: ✅ shipped · 🟡 partial · 🔜 planned. If a claim and a status disagree, the code wins — file an issue.

Evaluate This Repo

Before you fork, run this prompt against the codebase with any AI coding assistant:

You are a cynical senior engineer evaluating whether to fork this repo as the
base for a bots-that-run-workflows product. No prior context — explore it
as you naturally would. Tell me: should I fork this, and what's your honest
take? Be specific: file paths, line numbers, actual evidence. "The docs are
bad" is useless. Under 500 words.

We run this ourselves before every release. If the AI finds something we missed, file an issue.

The Name

Grok is named for a verb that means to understand something completely and intuitively. That is a confident name. It promises you a machine that gets it.

gawkbot is named for a verb that means to stare at something openly, at length, and without taking in any of it. That is the honest name for what actually happens once the bots are running. The work is done. You are looking at a dashboard about it.

We are a parody of a bot company. We also ship, which is the part they did not see coming.

Star History