Data & API

Every number on the dashboard is reproducible from the source. Pull it as CSV or JSON, no key required.

Open by default

The Dead Internet Monitor publishes its aggregate measurements as open data. The same daily readings that drive the dashboard, the Autopsy Matrix, and the headline Dead percentage are available through three read-only endpoints below. DIM-authored fields — classifications, confidence scores, bot scores, and every derived metric — are released under CC BY 4.0. Quote them, chart them, build on them; just attribute DIM.

No authentication, no rate-limit headers, no sign-up. Responses set Access-Control-Allow-Origin: * so they can be fetched from a browser. Endpoints are cached at the edge and refresh after each night’s aggregation.

Aggregate time-series

GET /api/export/aggregate

One row per day per source, plus a combined internet-wide row — the full daily measurement series behind the headline metrics. This is the endpoint for trend analysis, replication, and citation.

Parameters

Parameter	Meaning
`source`	Restrict to one platform (e.g. `reddit`, `hackernews`, `bluesky`). Omit for all sources plus the combined row.
`start_date`	Inclusive lower bound, `YYYY-MM-DD` (UTC).
`end_date`	Inclusive upper bound, `YYYY-MM-DD` (UTC).
`format`	`csv` (default) or `json`. JSON wraps the rows in a `meta` block carrying the licence and row count.

Columns

Column	Meaning
`date`	UTC date of the reading.
`source`	Platform, or `combined` for the internet-wide row.
`total_collected`	Every item collected that day — the coverage denominator. Coverage = `total_classified` / `total_collected`; the difference is short content skipped for too little signal.
`total_classified`	Items classified that day (the AI-rate denominator).
`ai_generated_count` / `human_count` / `uncertain_count`	Classification tallies.
`ai_percentage`	Daily AI-generated share (%).
`ai_percentage_30d`	30-day rolling AI share — the headline-stable figure.
`ai_percentage_ex_automod` / `_30d`	AI share excluding Reddit AutoModerator posts.
`automod_count`	AutoModerator items in the denominator.
`ai_percentage_ex_orgsync` / `_30d`	AI share excluding organisational-syndication (news-org handles).
`orgsync_count`	Org-syndication items.
`dead_internet_index` / `_30d`	Composite DII (supplementary technical index).
`dead_pct_30d`	30-day Dead-quadrant share — the headline metric.
`alive_count` / `zombified_count` / `polluted_count` / `dead_count`	Autopsy Matrix quadrant tallies.
`bot_engagement_pct`	Share of engagement from bot-scored authors (sources with an engagement signal).
`avg_confidence`	Mean classifier confidence.

Examples

All sources, CSV: /api/export/aggregate
Reddit only, JSON: /api/export/aggregate?source=reddit&format=json
Combined row, June 2026, CSV: /api/export/aggregate?source=combined&start_date=2026-06-01&end_date=2026-06-30

Item-level export

GET /api/export

Individual classified items with their DIM-authored fields — classification, confidence, bot score, and content-type tags. Use it to audit how a headline number was reached, item by item.

Two licences in one file. DIM-authored fields are CC BY 4.0. The title, author, source_url, and original post text remain the copyright of their authors and the terms of the originating platform — they are not covered by CC BY 4.0. The JSON meta.license block spells this out on every response.

Parameters

Parameter	Meaning
`source`	Restrict to one platform.
`classification`	`ai_generated`, `human`, or `uncertain`.
`min_confidence`	Float 0–1; drop items below this classifier confidence.
`content_class`	Keep only a tag, e.g. `automated_moderation`, `organizational_syndication`.
`exclude_content_class`	Drop a tag — e.g. exclude AutoModerator.
`start_date` / `end_date`	`YYYY-MM-DD` (UTC), inclusive.
`format`	`csv` (default) or `json`.

Examples

High-confidence AI items, JSON: /api/export?classification=ai_generated&min_confidence=0.9&format=json
Reddit, excluding AutoModerator, CSV: /api/export?source=reddit&exclude_content_class=automated_moderation

Live snapshot

GET /api/stats

The current headline state as JSON — the same object the dashboard renders: overall AI percentage, Dead quadrant, DII, Autopsy Matrix, and per-source summary. Accepts optional source, start_date, and end_date.

Current snapshot: /api/stats

Other ways to pull the number

Embeddable badges. Drop a live SVG number into an article, README, or dashboard — for the whole internet, any platform, or any community. See the badge gallery.
RSS. Follow the daily reading and new findings at /rss.xml.
Dataset metadata. The homepage carries schema.org Dataset markup for Google Dataset Search.

Licence & citation

DIM-authored aggregate data is licensed Creative Commons Attribution 4.0 (CC BY 4.0). You may share and adapt it, including commercially, with attribution. The required attribution string is:

Source: Dead Internet Monitor — deadinternetmonitor.com (CC BY 4.0)

Attribution for reused data.

For academic citation:

Dead Internet Monitor (2026). AI content and bot-activity
measurement data. deadinternetmonitor.com. Accessed: YYYY-MM-DD.

Suggested citation format.

Full licensing terms, including the boundary between DIM-authored fields and source content, are on the methodology page.

Building something with DIM’s data, or need a cut we don’t publish? hello@deadinternetmonitor.com.