Get a free trial key → · Full MCP reference · Or use Claude Desktop
https://app.staticowl.com/api/mcp
with no Authorization header — the anonymous tool catalog exposes a single
request_trial_key tool. Ask Claude Code to "sign me up for StaticOwl";
it collects your email + use case, fires the verification email, waits for you to click,
then registers itself with your key. Zero clicks in the browser after the verification link.
Step 1 — Get a free trial key (no card)
Visit app.staticowl.com/signup-key.html, drop your email + one line describing what you're building, submit. We email a verification link; click it, the page shows your so_* trial key + copy-paste blocks for every MCP client we support (including Claude Code).
- 30 days — auto-provisions a real sandbox site with its own graph + seeded content, ready to write to.
- Full MCP access against your sandbox — schema, content, media, builds, analytics, the works.
- Free — no card, no credit checks, no chasing.
- Same product — the trial API surface is identical to the paid one; the only difference is the sandbox scope.
Step 2 — Add StaticOwl to Claude Code
One command in your terminal:
claude mcp add staticowl \
--scope user \
--transport http \
--header "Authorization: Bearer YOUR_KEY_HERE" \
https://app.staticowl.com/api/mcp
That's it. claude mcp list should now show staticowl. The verified-key page shows this command with your key already filled in.
Prefer a project-scoped config?
If you want the MCP server available only inside one repository (say, your marketing site's checkout), drop this into .mcp.json at the repo root:
{
"mcpServers": {
"staticowl": {
"transport": {
"type": "http",
"url": "https://app.staticowl.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_KEY_HERE"
}
}
}
}
}
Commit .mcp.json (with the key rotated out for a placeholder) so teammates can wire up their own keys and share the config. Claude Code prompts before running project-scoped servers the first time — safer than user-scoped defaults.
Verify it worked
# List every MCP server Claude Code sees
claude mcp list
# Open a chat that has staticowl available + inspect the tool catalog
claude
> /mcp
The /mcp slash command shows every tool the server exposes, grouped by pack — sites, content, schema, templates, media, builds, chat, ops, insights, authoring, access. 98+ tools total.
Step 3 — Three prompts to try
"Show me my sandbox"
Use the StaticOwl MCP to list my sites. There should be one called "Sandbox" — show me its details (graph name, environments, theme) and tell me what content types it was seeded with. Then list the sample pages.
"Draft a blog post with a real image"
In my Sandbox, write a 500-word blog post about "why static site generators are having a renaissance." First call
voice_getto match the site's voice. Then generate a hero image with your image-gen tool, upload it to the site viamedia_upload_from_url, and create the blog post as a draft — with SEO metadata (viaseo_set) and 3 relevant tags (pick fromtags_listfirst). Show me a diff before you save each step.
"Plan a docs site end-to-end"
I want to extend my Sandbox to be a documentation site for an open-source library. Design new content types (Doc, Section, ApiReference), create sample content for a fictional graph-database library, wire up a template, and publish. Use your MCP tools for every step. When you hit permission edges, tell me. When you succeed, show me the URL.
What the tool catalog covers
| Pack | Tools | What Claude Code can do |
|---|---|---|
| Sites | 5 | List, get, create, update, set-default |
| Content | 10 | List, get, create, update, delete, publish, drafts, search, history, tags |
| Schema | 9 | Types, queries, scripts — list / get / create / update / delete |
| Templates | 7 | List, get, create, update, enable/disable, delete, preview |
| Media | 6 | List, get, upload from URL, generate alt text, delete, search stock |
| Authoring | 8 | Voice profile get/set, SEO get/set/inspect/social-preview, tags list/apply |
| Insights | 8 | Comments moderation, analytics, search analytics + zero-result queries |
| Chat | 6 | Conversation history + episodic memory (replay, consolidate, summarize) |
| Ops | 25 | Forms, redirects, workflows, feedback moderation, integrations config |
| Builds | 5 | Trigger, history, impact, last-status, rollback |
| Access | 6 | Users, roles, grant/revoke, audit chain verify, PII retention |
| Introspection | 2 | Views, routes |
For the full schema of every tool, run /mcp inside Claude Code, or read the reference at /docs/mcp/.
Why Claude Code + StaticOwl works
- Terminal-native workflow: You're already in your editor. Claude Code is right there. Ask it to draft an article, upload a hero image, publish — no context switch to a browser tab.
- Repo-scoped MCP config: Team members wire up their own keys via a shared
.mcp.json. Rotate one key, the others keep working. - Every operation is auditable: The bitemporal graph engine stamps every MCP call. "What did the AI do at 2:14pm last Tuesday" is a query, not a forensics project.
- Same authority as a human admin: Whatever permissions your key has, the agent has. Trial keys are sandbox-scoped; full-account keys aren't. Nothing sneaks through.
- HTTP-native, no bridge: Claude Code speaks streamable HTTP MCP transport. No
mcp-remote, no stdio contortion. Faster round-trips + real streaming responses.
Curl it directly (no client needed)
The whole MCP surface is HTTP-addressable. Useful when you're debugging or scripting outside Claude Code:
# Health check
curl https://app.staticowl.com/api/health
# Who am I? (verifies the key works)
curl -H "Authorization: Bearer $KEY" https://app.staticowl.com/api/access/me
# List every MCP tool the server exposes
curl -X POST https://app.staticowl.com/api/mcp \
-H "Authorization: Bearer $KEY" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'
# Call a tool — here, upload a generated image from a URL
curl -X POST https://app.staticowl.com/api/mcp \
-H "Authorization: Bearer $KEY" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{
"jsonrpc":"2.0","id":2,"method":"tools/call",
"params":{
"name":"media_upload_from_url",
"arguments":{"sourceUrl":"https://picsum.photos/1200","altText":"Random landscape"}
}
}'
FAQ
Does this cost money to demo?
No. Email verification only, no card, 30 days on the sandbox. The trial API surface is the same as the paid product; the sandbox is a real site with its own graph you can actually publish from. When the 30 days end, the sandbox is dropped unless you've subscribed — subscribing keeps everything you built.
What's the difference from Claude Desktop?
Same MCP server, same tool set. Claude Code is the terminal-native flavour — better fit for a developer already living in ~/projects/. Claude Desktop is the GUI flavour — better fit for content teams. Both work. You can have both configured against the same trial key.
What tools does Claude Code call by default?
Whatever your prompt implies. Claude Code inspects tools/list, sees the descriptions, and picks the right ones. Ask "publish this article" and it'll call content_publish after saving. Ask "what do visitors search for that we don't answer" and it'll call search_analytics_zero_results. The tool descriptions are written for this — no prompt engineering required.
Can I check into the repo without leaking my key?
Yes. Use the project-scoped .mcp.json with an env-var placeholder — Claude Code substitutes it at runtime. Add STATICOWL_KEY to your .env (gitignored) and reference it in .mcp.json via "Authorization": "Bearer $STATICOWL_KEY". Teammates set their own STATICOWL_KEY in their local env.
Is it really MCP-native?
Native. The MCP server is the same HTTP surface as the CMS API — every tool is an Express route under the hood, so authorization, RBAC, audit chain, and workflow gates all apply identically to human and agent traffic. There is no "MCP tier" — the tool catalog IS the product.
Get started
Free 30-day trial. No card. Email verification only. Real sandbox site with its own graph. Same product surface as the paid tier. Get my trial key →
Questions? Email support@staticowl.com.