Search the SiteGPT documentation: setup, features, API reference, troubleshooting.
io.github.sitegpt/sitegpt-docs MCP Server
This MCP server provides searchable access to the SiteGPT documentation, covering setup, features, API reference, and troubleshooting. It exposes 3 tools and is designed to support agent environments that need to discover and use SiteGPT guidance programmatically.
🛠️ Key Features
Search SiteGPT documentation content (setup, features, API reference, troubleshooting)
Provides agent-oriented documentation for “SiteGPT Agent Skills”
Includes an Auth.md discovery entry for try-before-signup onboarding via https://sitegpt.ai/auth.md
🚀 Use Cases
Enable agent environments to manage SiteGPT tasks using SiteGPT CLI
Support automation around chatbots, knowledge, conversations, settings, team access, and account resources
Provide onboarding discovery for anonymous try-before-signup flows
⚡ Developer Benefits
Tool access count: 3 tools
Documentation content structured for agent environments (skills and Auth.md-style discovery)
⚠️ Limitations
Documentation coverage is limited to what the server indexes: setup, features, API reference, and troubleshooting
Captured live from the server via tools/list.
search_site_gpt_docs
Search across the SiteGPT Docs knowledge base to find relevant information, code examples, API references, and guides. Use this tool when you need to answer questions about SiteGPT Docs, find specific documentation, understand how features work, or locate implementation details. The search returns contextual content with titles and direct links to the documentation pages. If you need the full content of a specific page, use the query_docs_filesystem tool to `head` or `cat` the page path (append `.mdx` to the path returned from search — e.g. `head -200 /api-reference/create-customer.mdx`).
Run a read-only shell-like query against a virtualized, in-memory filesystem rooted at `/` that contains ONLY the SiteGPT Docs documentation pages and OpenAPI specs. This is NOT a shell on any real machine — nothing runs on the user's computer, the server host, or any network. The filesystem is a sandbox backed by documentation chunks.
This is how you read documentation pages: there is no separate "get page" tool. To read a page, pass its `.mdx` path (e.g. `/quickstart.mdx`, `/api-reference/create-customer.mdx`) to `head` or `cat`. To search the docs with exact keyword or regex matches, use `rg`. To understand the docs structure, use `tree` or `ls`.
**Workflow:** Start with the search tool for broad or conceptual queries like "how to authenticate" or "rate limiting". Use this tool when you need exact keyword/regex matching, structural exploration, or to read the full content of a specific page by path.
Supported commands: rg (ripgrep), grep, find, tree, ls, cat, head, tail, stat, wc, sort, uniq, cut, sed, awk, jq, plus basic text utilities. No writes, no network, no process control. Run `--help` on any command for usage.
Each call is STATELESS: the working directory always resets to `/` and no shell variables, aliases, or history carry over between calls. If you need to operate in a subdirectory, chain commands in one call with `&&` or pass absolute paths (e.g., `cd /api-reference && ls` or `ls /api-reference`). Do NOT assume that `cd` in one call affects the next call.
Examples:
- `tree / -L 2` — see the top-level directory layout
- `rg -il "rate limit" /` — find all files mentioning "rate limit"
- `rg -C 3 "apiKey" /api-reference/` — show matches with 3 lines of context around each hit
- `head -80 /quickstart.mdx` — read the top 80 lines of a specific page
- `head -80 /quickstart.mdx /installation.mdx /guides/first-deploy.mdx` — read multiple pages in one call
- `cat /api-reference/create-customer.mdx` — read a full page when you need everything
- `cat /openapi/spec.json | jq '.paths | keys'` — list OpenAPI endpoints
Output is truncated to 30KB per call. Prefer targeted `rg -C` or `head -N` over broad `cat` on large files. To read only the relevant sections of a large file, use `rg -C 3 "pattern" /path/file.mdx`. Batch multiple file reads into a single `head` or `cat` call whenever possible.
When referencing pages in your response to the user, convert filesystem paths to URL paths by removing the `.mdx` extension. For example, `/quickstart.mdx` becomes `/quickstart` and `/api-reference/overview.mdx` becomes `/api-reference/overview`.
Parameters1
command
string
required
A shell command to run against the virtualized documentation filesystem (e.g., `rg -il "keyword" /`, `tree / -L 2`, `head -80 /path/file.mdx`).
Raw schema
{
"type": "object",
"properties": {
"command": {
"type": "string",
"description": "A shell command to run against the virtualized documentation filesystem (e.g., `rg -il \"keyword\" /`, `tree / -L 2`, `head -80 /path/file.mdx`)."
}
},
"required": [
"command"
],
"additionalProperties": false
}
submit_feedback
Report a problem with this documentation site so the docs team can fix it. Use when a documentation page is incorrect, outdated, confusing, incomplete, or has a broken example. This is for feedback about the documentation content itself — not for product support requests or feedback about this tool or assistant.
Parameters2
path
string
required
The documentation page path the feedback is about (e.g., the page you were reading, such as `/quickstart`).
feedback
string
required
A clear description of the documentation issue or suggestion — what is incorrect, outdated, missing, or confusing.
Raw schema
{
"type": "object",
"properties": {
"path": {
"type": "string",
"minLength": 1,
"description": "The documentation page path the feedback is about (e.g., the page you were reading, such as `/quickstart`)."
},
"feedback": {
"type": "string",
"minLength": 1,
"description": "A clear description of the documentation issue or suggestion — what is incorrect, outdated, missing, or confusing."
}
},
"required": [
"path",
"feedback"
],
"additionalProperties": false
}
These skills teach agents how to use SiteGPT through the SiteGPT CLI so they can manage chatbots, knowledge, conversations, settings, team access, and account resources from an agent environment.
SiteGPT also publishes https://sitegpt.ai/auth.md for agent environments that
support Auth.md-style discovery of the anonymous try-before-signup onboarding
flow.
With the skills CLI (any agent environment that supports Agent Skills):
code
npx skills add sitegpt/agent-skills
Both deliver the same skill. This repository is dual-packaged: an Agent
Skills repo (skills.sh) and a Claude Code plugin marketplace
(.claude-plugin/).
Available Skills
sitegpt
Orientation skill: what SiteGPT is, which surface to use (CLI, MCP
connector, REST API, or no-signup onboarding), and where the deep
workflows live. Routes terminal work to sitegpt-cli.
sitegpt-cli
Use the SiteGPT CLI to manage SiteGPT accounts and chatbots from AI agents such as Codex, Claude Code, Cursor, OpenCode, Gemini CLI, Windsurf, Cline, and other skill-compatible coding agents.
Use when:
Creating a SiteGPT chatbot from a website.
Creating a try-before-signup chatbot through agent-first onboarding when the
user does not have a SiteGPT account yet.
Adding knowledge from links, websites, sitemaps, files, YouTube videos, text, and connected data sources.
Managing personas, instructions, settings, conversation starters, followups, and custom responses.
Reading and managing conversations, messages, leads, tags, members, invites, usage, billing, and API tokens.
Troubleshooting SiteGPT CLI command usage.
The skill is a single self-contained SKILL.md: the agent-facing workflow,
discovery brief, onboarding and account playbooks, command map, and safety
rules in one file. The SiteGPT CLI itself is the source of truth for exact
command syntax — agents run sitegpt <command> --help for flags, so the skill
deliberately does not duplicate a per-command reference.
Install in Cursor
This repository is also a Cursor plugin (skills + the remote MCP server).
In Cursor, run /add-plugin and pick SiteGPT, or install from the
marketplace listing. The MCP server connects your SiteGPT account with
browser OAuth; the skills work with no account via agent-first onboarding.
Installation
Install the SiteGPT CLI skill with the open skills CLI:
The skills CLI installs this agent skill. The SiteGPT CLI is the actual command-line tool the agent will run after the skill is installed.
Install the SiteGPT CLI first:
bash
npm install -g @sitegpt/cli
Then choose the right flow:
No SiteGPT account yet: do not log in first. Ask the agent to run
sitegpt onboarding start <website-url>, configure and test the temporary
chatbot, then share the onboarding URL for preview and claim.
Existing SiteGPT account: authenticate, then use normal account commands.
For existing accounts, authenticate with device login:
bash
sitegpt login
You can also create an API token from the SiteGPT dashboard and save it manually:
bash
sitegpt login --token <sitegpt-api-token>
Example Prompts
Once the skill is installed, ask your agent:
text
Try SiteGPT for https://example.com. Inspect the website, create a temporary chatbot, add knowledge, configure persona and instructions, test it, and give me the onboarding URL so I can preview and claim it.
text
I already have a SiteGPT account. Create a chatbot for https://example.com inside my account, configure knowledge and branding, and give me the dashboard link.
text
Audit my SiteGPT chatbot knowledge sources and resync any failed or stale documents.
text
Show recent conversations for my SiteGPT chatbot, summarize unresolved issues, and tag the conversations that need human follow-up.
Skill Structure
text
skills/
sitegpt-cli/
SKILL.md
Versioning And Releases
Skill versions live in metadata.version inside each SKILL.md and in the root package.json.
Releases are manual. When publishing a new version:
Update package.json.
Update metadata.version in each changed SKILL.md.
Update CHANGELOG.md.
Commit with chore: release <version>.
Create and push a matching Git tag, for example v0.1.2.
Create a GitHub Release for that tag.
Users who install through npx skills add sitegpt/agent-skills can update later with npx skills update.