Content Tools

sourcemap.news

Provides daily news story editions with ranked summaries, fact receipts, source links, and verification details.

1 endpoint29 known toolsFirst detected October 22, 2026Last detected October 22, 2026

ENDPOINT 1

https://sourcemap.news/mcp

No auth detected

MCP server metadata

Name: source-map-agent-session
Version: 0.5.0

Capabilities

tools

Server instructions

Source Map: AI news, traced to where it started. This is the edition of 9 October 2026, with the earlier editions standing behind it in the same world (a story id is '<edition date>/<name>'; list_stories and get_story read any edition). The website is a 3D map with one place per story. The stories are numbered in edition order (by section, then position), which is not a ranking by score: for the highest scoring first, call list_stories with order "score". Read with describe_edition, whats_new, list_stories, get_story and check_receipt (it never fetches for you); send a find, evidence or a correction with submit_find (stored as data for review). To show a person, call start_session and give them the join_url; fly_to, open_receipt, highlight_door, start_tour and show_map take them through it. Only the person opens source links.

Known tools 29

describe_edition

Start here.

Inferred read-only

whats_new

What has changed since you last looked, across every day.

Inferred read-only

list_stories

The stories with their scores out of 9 (newness, momentum, evidence, each 0 to 3) and rank by score, their section, their place on the map (deep link #story=<id>), and how many facts (SAW, STATED and INFER) and sources each has.

Inferred read-only

get_story

One story with its receipts as it stands now: summary, who started it, how we found it, why it matters, its score, every fact with its mark (SAW: fetched and read by us; STATED: a source's own claim, quoted; INFER: our reasoning, with the lines it rests on, never fetched), who stated it, when a number was read, and the source links, each receipt added since it was first published with when it was added, when it was first published (and in which edition) and last updated, and its own page.

Inferred read-only

check_receipt

One fact's stored receipt (its words, its mark, who stated it, when it was read, or for an INFER line the lines it rests on) and the story's source links, with how to check it.

Inferred read-only

get_scores

Every story's score out of 9, ranked highest first, with the momentum actually measured and the rubric (printed in the world so anyone can check a score by hand).

Inferred read-only

find_stories

Find stories on any day.

Inferred read-only

submit_find

Send us a find (something new and small in AI, traced as far back as you can), evidence for one of this edition's stories, or a correction, with at least one receipt (a link, and the exact words you read there, and when).

Potential side effects

list_items

v2's list: the stories in order with id, number, section, headline, summary, score total and counts.

Inferred read-only

get_item

v2's name for get_story.

Inferred read-only

start_session

Open an agent session for one person.

Inferred read-only

wait_for_events

Long poll: returns as soon as the PERSON does something worth a response (or when max_wait_s passes): joined, arrived (they settled at a place), receipt_opened, receipt_closed, source_opened (they opened a source door), map, playback (tour), took_control, view_changed, session_ended.

Inferred read-only

fly_to

Fly the person's world to a place: a story id ('<edition date>/<name>'), an edition's date (the front of that edition), 'arrival' (the front of the current edition), 'how', 'earlier' (the sign behind today's front that lists the earlier editions), or 'end'.

Inferred read-only

open_receipt

Fly to a story and open its receipt: every fact stamped SAW, STATED or INFER, who stated it, read times.

Inferred read-only

close_receipt

Close the open receipt.

Inferred read-only

highlight_door

Make one of a story's source doors beckon (source is 0 based, in the story's order), with your reason as a caption and the door's destination shown.

Inferred read-only

start_tour

Start (or resume) the current edition's guided tour: the narration (about 4 minutes) with the camera in step, each narrated story's receipt printing as its facts are spoken.

Inferred read-only

play_story

Play ONE story's narration on the person's page, then stop: the camera goes to that story and its receipt prints in step as its facts are said (a story's clip is about 20 to 40 seconds, said in our own voice from the story's record only).

Inferred read-only

pause_tour

Pause the tour where it is (the person can look around, then resume).

Inferred read-only

show_map

Open (open: true, the default) or close the map: the world from above, every story's traced route on the ground, and every place of every edition listed with its score, newest edition first.

Inferred read-only

get_position

Where the person is now: the place (story id, n and headline, or arrival, how, earlier, end, or an edition's date), the edition they are in, whether a receipt or the map is open, a highlighted door, the tour, the view (world or text) and the detail (full or light), and who moved last (person or agent).

Inferred read-only

open_item

v2's name for fly_to with a story id.

Inferred read-only

show_receipts

v2's name: fly to a story and open its receipt with every line printed.

Inferred read-only

play

v2's name for start_tour (resume from where the tour is).

Inferred read-only

pause

v2's name for pause_tour.

Inferred read-only

seek_story

Move the tour to a story's stop (the camera lands there, in step with the narration).

Inferred read-only

act

The general form of the session tools above.

Inferred read-only

get_state

v2's name for get_position (the page's latest state).

Inferred read-only

end_session

End the session; the page tells the person their assistant has left.

Inferred read-only

CONNECT WITH APPROVAL

Client installation

Review this server and its permissions before adding it. Secret placeholders must be set locally.

Codex

~/.codex/config.toml

[mcp_servers.source-map-agent-session]
url = "https://sourcemap.news/mcp"
enabled = true

Claude Code

.mcp.json

{
  "mcpServers": {
    "source-map-agent-session": {
      "type": "http",
      "url": "https://sourcemap.news/mcp"
    }
  }
}

Claude Desktop

Settings → Connectors → Add custom connector

Name: source-map-agent-session
Remote MCP URL: https://sourcemap.news/mcp

Add this remote URL as a custom connector in Claude Desktop. Availability depends on the user plan and workspace policy.

Cursor

.cursor/mcp.json

{
  "mcpServers": {
    "source-map-agent-session": {
      "url": "https://sourcemap.news/mcp"
    }
  }
}

Visual Studio Code

.vscode/mcp.json

Add to Visual Studio Code

{
  "servers": {
    "source-map-agent-session": {
      "type": "http",
      "url": "https://sourcemap.news/mcp"
    }
  }
}

Generic MCP

Client-specific MCP configuration

{
  "name": "source-map-agent-session",
  "transport": "streamable-http",
  "url": "https://sourcemap.news/mcp"
}

MCP Inspector

Run the official MCP Inspector locally and enter the indexed Streamable HTTP endpoint.

TRUST AND VERIFICATION EVIDENCE

Loading Trust v2 evidence…

Checking the associated registrable domain. The BuiltWith key remains server-side.

Indexed

Evidence is source-attributed and does not guarantee that a third-party server is safe. Risk labels are conservative metadata heuristics.