Proofite MCP server
{
"mcpServers": {
"proofite": {
"command": "npx",
"args": [
"-y",
"proofite-mcp@latest"
],
"env": {
"PROOFITE_API_KEY": "<proofite_api_key>"
}
}
}
}{
"mcpServers": {
"proofite": {
"command": "npx",
"args": [
"-y",
"proofite-mcp@latest"
],
"env": {
"PROOFITE_API_KEY": "<proofite_api_key>"
}
}
}
}{
"mcpServers": {
"proofite": {
"command": "npx",
"args": [
"-y",
"proofite-mcp@latest"
],
"env": {
"PROOFITE_API_KEY": "<proofite_api_key>"
}
}
}
}{
"servers": {
"proofite": {
"command": "npx",
"args": [
"-y",
"proofite-mcp@latest"
],
"env": {
"PROOFITE_API_KEY": "${input:proofite_api_key}"
}
}
},
"inputs": [
{
"id": "proofite_api_key",
"type": "promptString",
"description": "Proofite api key",
"password": true
}
]
}{
"mcpServers": {
"proofite": {
"command": "npx",
"args": [
"-y",
"proofite-mcp@latest"
],
"env": {
"PROOFITE_API_KEY": "<proofite_api_key>"
}
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall npx -y proofite-mcpversion 1.0.0
What it does
Proofite is an MCP server that connects to a personal AI-written news briefing service. It lets agents read daily briefings and adjust the information filter by tuning topics, correcting briefings with feedback, and managing sources like newsletters, RSS feeds, and web searches.
- Read your daily AI-written briefing and see which stories were selected from your subscribed newsletters and feeds.
- Tell your agent 'too much crypto this week, go deeper on EU regulation' and have it adjust topic weights and queue research for tomorrow's briefing.
- Inspect your information diet settings, manage topic preferences, and request deep-dive research on specific subjects.
Tools this server exposes
| Tool | Does |
|---|---|
| get_briefing | Read the user's personal AI-written daily briefing in markdown with all stories, sources, and entry ids for rating. |
| list_briefings | List past daily briefings with date, title, story count, and podcast availability. |
| get_information_diet | Read the full picture of what shapes the briefing: schedule, length, technicality, depth, interests, no-go topics, and learned preferences. |
| list_sources | List all topic inboxes feeding the briefing, including newsletters, RSS feeds, and monitored web searches with recent activity. |
| get_source | Inspect one topic inbox in detail: RSS feeds, web searches, X/Twitter profiles, and newsletters. |
| get_podcast_episode | Get the audio episode generated from a briefing with a playable URL and show notes listing sources. |
| list_read_later | List links and forwarded emails queued for the next briefing. |
| list_deep_dives | List deep-dive research requests waiting to be covered in future briefings. |
| get_account | Check account status: plan, briefing schedule, source and story counts, and current AI costs. |
| update_information_diet | Change global briefing settings: delivery time, length, technicality, depth, stance, tone, language, and podcast preferences. |
| tune_topic | Adjust how a single topic is treated: frequency, technicality, depth, or mute it entirely. |
| correct_briefing | Give plain-language feedback on a briefing and have an AI editor convert it into concrete topic and setting adjustments. |
| rate_briefing_entry | Signal approval or criticism of one story with options like 'more technical', 'follow it', or 'mute this topic'. |
| set_learned_profile | Read or update the free-text profile describing the user's work, projects, and interests that the briefing uses. |
| request_deep_dive | Request the next briefing to research a topic with fresh web searches and include a mini-dossier. |
| save_for_later | Add a link or forwarded email to the read-later inbox for coverage in the next briefing. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.