About Entry
Entry is a remote MCP server for connecting AI assistants to personal profiles and writing drafts. It supports reading and editing drafts and proposing profile changes, while public changes require the owner’s approval in Entry.
Key features
- 01Read profiles, drafts, and posts
- 02Create and edit private drafts
- 03Propose profile changes for approval
- 04Request publication of a draft
- 05Track publication request status
Use cases
- 01Find unfinished drafts and review their content
- 02Prepare an article from an idea
- 03Suggest profile edits without changing the live page
- 04Submit a finished draft for owner approval
MCP support
Entry exposes profile, draft, and publication-request tools through a remote Streamable HTTP MCP server at `https://entry.fyi/mcp`. It uses OAuth; users select which Entries and permissions to share, and approve public changes in Entry.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://entry.fyi/mcpclaude mcp add --transport http fyi-entry-entry https://entry.fyi/mcp{"mcpServers":{"fyi-entry-entry":{"type":"http","url":"https://entry.fyi/mcp"}}}Open in Cursor{"servers":{"fyi-entry-entry":{"type":"http","url":"https://entry.fyi/mcp"}}}Open in VS Code{"mcpServers":{"fyi-entry-entry":{"type":"http","url":"https://entry.fyi/mcp"}}}{"mcpServers":{"fyi-entry-entry":{"serverUrl":"https://entry.fyi/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEEntry MCP server
Connect your AI assistant to Entry, your personal page and your writing. The assistant can read your profile and drafts, write and edit drafts, propose profile changes and ask you to publish. You approve every public change in Entry.
https://entry.fyi/mcp
A remote MCP server: Streamable HTTP, signed in with OAuth. There's nothing to install or run, and no API key. When your client connects, Entry asks you to sign in and choose which of your Entries to share and what the assistant may do.
- Connect your assistant
- What the assistant can do
- Tools
- Permissions
- For client developers
- Agent Skills
Connect your assistant
Claude Code
claude mcp add --transport http entry https://entry.fyi/mcp
Then, inside Claude Code, run /mcp and sign in to Entry in your browser. Add --scope user to use it in every project.
Or install the plugin, which adds the server and the Entry skills:
/plugin marketplace add entry-fyi/skills
/plugin install entry@entry
Claude (claude.ai and desktop)
- Open Customize → Connectors and choose Add custom connector. On Team and Enterprise plans an Owner adds it under Organization settings → Connectors → Add → Custom, and members then press Connect.
- Name it Entry and paste
https://entry.fyi/mcp. - Choose Add. Claude sends you to Entry: sign in, tick the Entries to share and choose Allow.
- In a new chat press +, then Connectors, and turn Entry on.
ChatGPT
- Open ChatGPT Plugins. If your plan needs developer mode, turn it on first under Settings → Apps → Advanced.
- Select +, then Add custom MCP server. Name it Entry.
- Under Connection, paste
https://entry.fyi/mcp. For authentication choose OAuth and leave any client ID and secret empty. - Read the risk warning, confirm, and create it.
- ChatGPT sends you to Entry. Sign in, tick the Entries to share and choose Allow. Start a new chat and turn Entry on for it.
Cursor
Or add it to ~/.cursor/mcp.json:
{
"mcpServers": {
"entry": { "url": "https://entry.fyi/mcp" }
}
}
VS Code
Add to VS Code · Add to VS Code Insiders
Or add it to .vscode/mcp.json:
{
"servers": {
"entry": { "type": "http", "url": "https://entry.fyi/mcp" }
}
}
Codex CLI
codex mcp add entry --url https://entry.fyi/mcp
codex mcp login entry
Or add it to ~/.codex/config.toml:
[mcp_servers.entry]
url = "https://entry.fyi/mcp"
Gemini CLI
gemini extensions install https://github.com/entry-fyi/skills
That adds the server and the Entry skills. Or add just the server:
gemini mcp add --transport http entry https://entry.fyi/mcp
Then, inside the Gemini CLI, run /mcp auth entry and sign in in your browser.
Antigravity CLI
Add it to ~/.gemini/config/mcp_config.json:
{
"mcpServers": {
"entry": { "serverUrl": "https://entry.fyi/mcp" }
}
}
Zed
Add it to your settings.json:
{
"context_servers": {
"entry": { "url": "https://entry.fyi/mcp" }
}
}
Grok
- Open Grok connectors. On Grok Business and Enterprise a team admin adds Entry first, in the xAI console under Grok Business → Connectors → Add Connector → Other.
- Choose New Connector, then Custom. Name it Entry and paste
https://entry.fyi/mcpas the Server URL. - Choose Add Connector. Grok sends you to Entry: sign in, tick the Entries to share and choose Allow.
Any other MCP client
Entry is a standard remote server, so any client that supports remote servers with OAuth works. A typical configuration:
{
"mcpServers": {
"entry": { "type": "http", "url": "https://entry.fyi/mcp" }
}
}
Where your client wants a key other than url, the URL goes in the same place.
The same steps, kept current, are at entry.fyi/developers#connect.
What the assistant can do
Try asking:
- "Show my unfinished Entry drafts."
- "Create an article about my recent work."
- "Update the introduction of this draft."
- "Prepare this article for publication."
- "Change my headline to …"
It can:
- Read your Entry as you see it, including fields you haven't published, and your drafts and posts.
- Start a draft from your idea and edit it with you. Drafts are private until you publish them.
- Propose changes to your profile: name, headline, bio, location, links, availability.
- Ask you to publish a draft when it's ready.
It can't publish, unpublish or change your live page. A profile proposal or a publication request appears in Entry, where you see exactly what would change and approve or reject it. In ChatGPT and Claude, profiles, drafts and publication requests show as cards in the chat.
Tools
| Tool | What it does | Writes? |
|---|---|---|
list_entries | Lists the Entries this connection can reach | no |
get_entry_profile | Reads a profile as the owner sees it, including unpublished fields | no |
list_drafts | Lists working or published drafts, 20 at a time | no |
get_draft | Reads one draft in full, with its revision | no |
create_draft | Creates a private draft | adds a draft |
update_draft | Saves changes to a draft against the revision last read | overwrites the fields sent |
propose_profile_update | Proposes profile changes for the owner to approve | adds a proposal |
request_publication | Asks the owner to publish a draft | adds a request |
get_publication_status | Checks a publication request: pending, published, rejected, stale or withdrawn | no |
Every tool has an input and output schema with Entry's field limits (maxLength, maxItems) and explicit annotations (readOnlyHint, destructiveHint, idempotentHint, openWorldHint). Full schemas: entry.fyi/developers#mcp-tools.
Permissions
When you connect, you choose which Entries the assistant may reach and which of five permissions it gets:
| Scope | Lets the assistant |
|---|---|
profile:read | Read your profile |
drafts:read | Read your drafts and posts |
drafts:write | Create and edit drafts |
profile:propose | Propose profile changes |
publish:request | Ask you to publish a draft |
Disconnect any time from Settings → Integrations → Connected apps in Entry; access stops on the next request. Entry shares your content with an assistant only when you ask it to use Entry, and doesn't use it to train AI models. See the privacy policy.
For client developers
| Endpoint | POST https://entry.fyi/mcp |
| Transport | Streamable HTTP, stateless, JSON responses (no sessions, no SSE). GET and DELETE are 405 |
| Auth | OAuth 2.1: authorization code with PKCE S256, dynamic client registration or a Client ID Metadata Document, resource=https://entry.fyi/mcp (RFC 8707) |
| Discovery | 401 with WWW-Authenticate: Bearer resource_metadata="https://entry.fyi/.well-known/oauth-protected-resource/mcp" |
| Authorization server | https://entry.fyi/.well-known/oauth-authorization-server |
| Personal access tokens | Authorization: Bearer entry_pat_…, made in Settings → Integrations, for testing with MCP Inspector |
| Limits | 120 reads and 30 writes a minute per connection; refusals carry Retry-After or details.retryAfter |
| MCP Apps | Cards for profiles, drafts and publication requests (ui://entry/*) in hosts that support them |
Check the server answers (without a token it must be 401 with a WWW-Authenticate header, which is how a client finds Entry's sign-in):
curl -i -X POST https://entry.fyi/mcp
Tool errors are tool results with isError: true and { code, message, details }: forbidden, not_found, conflict, validation_failed, limit_reached, rate_limited, bad_request.
Entry also has a public, read-only API for published profiles (no auth) and an authenticated REST API for the same operations as this server: entry.fyi/developers, OpenAPI.
Agent Skills
entry-fyi/skills has Agent Skills that teach an agent to use Entry well: entry (this server, the right tool for each request, revisions and the approval boundary) and entry-public-profiles (looking people and companies up through the public API).
npx skills add entry-fyi/skills
Support
- Help for owners: entry.fyi/developers/support
- Bugs and requests: open an issue or email [email protected]
- Terms · Privacy
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.