Every daily report and every library document is exposed over MCP, so Claude, ChatGPT, Grok or your own agent can query them directly — no scraping, no copy-paste, no stale context. One endpoint, eight tools.
Access is issued per client — requests open soon.
llm_guidance block telling the model how to read it — what the numbers mean, and what not to over-claim from them. Eight tools over two data sets — the daily report pipelines and the research library.
| Tool | Arguments | What it returns |
|---|---|---|
list_report_domains | — | List every report domain available. Call this first when unsure which domain to use. |
list_reports | domain, network?, limit? | Recent reports for one domain, newest first, as lightweight summaries. |
get_report | domain, network?, date | Full report for one domain on a specific UTC date, including its llm_guidance block. |
get_latest_report | domain, network? | Most recent report for one domain, same full content as get_report. |
| Tool | Arguments | What it returns |
|---|---|---|
list_library_documents | — | Every analyzed document in the research corpus — id, title, date, novelty score, digest headline. |
get_library_document | id | The full three-pass analysis (critique, report, digest) for one corpus document. |
get_library_synthesis | — | The single holistic synthesis across the whole corpus — cross-document claims, contradictions, open risks. |
get_library_context_pack | — | Curated glossary, canonical facts and guardrails, meant to be attached to an LLM session. |
The domain argument on the four report tools accepts these five keys.
| Key | Pipeline | Extra argument |
|---|---|---|
network | Network Health | requires network: mainnet | testnet |
social | X (Twitter) Activity | — |
bsc | BSC On-Chain Activity | — |
community | Telegram Community | — |
price | Price & Market | — |
Two ways in, depending on what your client supports. Both are unlocked by the same wallet sign-in.
Authorization: Bearer <token> on every request. Simplest path — used by mcp-remote bridges and clients that allow custom headers. The endpoint speaks Streamable HTTP, so anything MCP-compatible can connect.
In your client's connector settings, add a custom/remote MCP server and paste the endpoint URL. The client discovers the OAuth flow automatically and walks you through the consent screen.
https://mcp.telnetx.com/mcp as the server URL. Leave the transport on the default (Streamable HTTP). For clients that only launch local stdio servers, bridge with mcp-remote and pass the bearer token as a header.
{
"mcpServers": {
"netx-daily-reports": {
"command": "npx",
"args": [
"-y", "mcp-remote",
"https://mcp.telnetx.com/mcp",
"--header", "Authorization:Bearer ${MCP_TOKEN}"
],
"env": { "MCP_TOKEN": "<your token>" }
}
}
}The endpoint is a plain POST. List the tools to check your credentials work:
curl -X POST https://mcp.telnetx.com/mcp \
-H "Authorization: Bearer $MCP_TOKEN" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'…then call one:
{
"jsonrpc": "2.0",
"id": 2,
"method": "tools/call",
"params": {
"name": "get_latest_report",
"arguments": { "domain": "network", "network": "mainnet" }
}
}