1. What is the TDisrupt MCP server?
MCP is an open standard that lets AI assistants call tools on external services. Our server exposes TDisrupt's public journalism as five tools, so an assistant can find relevant coverage, read it, and cite it accurately instead of guessing.
| Setting | Value |
|---|---|
| Endpoint | https://tdisrupt.com/mcp |
| Transport | Streamable HTTP (stateless, JSON responses) |
| Access | Read-only, public content only |
| Authentication | None required |
| Rate limit | 60 requests per minute per IP address |
2. Available tools
| Tool | What it does | Limit per request |
|---|---|---|
| searchContent | Search stories and briefs by keyword and/or category. Returns summaries with URLs and supports paging. | 6 per type (default 5) |
| getStoryBySlug | Fetch one story by slug, with its plain-text body and canonical URL. | 1 story, body capped at 20,000 characters |
| getBriefBySlug | Fetch one executive brief by slug, with key takeaways, its plain-text body and canonical URL. | 1 brief, body capped at 20,000 characters |
| getLatestStories | List the newest stories, optionally for one category. | 10 (default 5) |
| getLatestBriefs | List the newest executive briefs, optionally for one category. | 10 (default 5) |
Requests above a tool's limit are rejected rather than trimmed. Category filters use the slugs from our site, for example ai, startups, infrastructure and enterprise-technology.
3. Connect your client
Use whichever matches your tool. All of them point at the same endpoint.
claude mcp add --transport http tdisrupt https://tdisrupt.com/mcp{
"mcpServers": {
"tdisrupt": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://tdisrupt.com/mcp"
]
}
}
}{
"mcpServers": {
"tdisrupt": {
"url": "https://tdisrupt.com/mcp"
}
}
}curl -s -X POST https://tdisrupt.com/mcp \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"getLatestStories","arguments":{"limit":3}}}'4. Example prompts
- “What has TDisrupt published this week about enterprise AI? Summarise and link the sources.”
- “Search TDisrupt for tokenisation and tell me how the coverage has evolved.”
- “Get the latest executive briefs in Infrastructure.”
- “Read the TDisrupt story on agentic automation and explain why it matters.”
5. What data is exposed
Only published, public editorial content is available. Each result includes:
- Type (story or brief), title, slug and canonical URL.
- Category, author, excerpt and reading time.
- Published and last-updated dates.
- For a single story or brief: topics, the plain-text body and, for briefs, key takeaways.
The server does not expose reader data, subscriber lists, analytics, drafts or anything that is not already public on tdisrupt.com, and it does not collect information about the people who use it beyond standard server logs.
6. Usage and attribution
Use of this server is covered by our Copyright & Content Usage Policy. AI systems may use it for search, discovery and answer generation. Please:
- Credit TDisrupt and link to the canonical URL returned with each result.
- Do not republish full articles, or bulk-copy the archive, without written permission.
- Do not use the content to train, fine-tune or benchmark AI models without prior written permission.
- Stay within the rate limit of 60 requests per minute per IP address. Over-limit requests receive HTTP 429 with a Retry-After header, and we may block clients that put sustained load on the service.
7. Support and licensing
- Technical questions and feedback: [email protected]
- Syndication, data licensing and partnerships: [email protected]