Broadcasting data, delivered clean. · Head-to-head
TVmaze API - TV Show and Episode Database vs Internet Archive TV News Archive
Which broadcasting data, delivered clean. data fits your job: TVmaze API - TV Show and Episode Database, or Internet Archive TV News Archive. API, files, or your warehouse. Daily, weekly, or hourly.
TVmaze API - TV Show and Episode Database
Internet Archive TV News Archive
Coverage, side by side
| TVmaze API - TV Show and Episode Database | Internet Archive TV News Archive | |
|---|---|---|
| Granularity | Show, cascading to season, episode, person and credit level | One broadcast item - typically a half-hour to hour-long programme segment |
What each contains
Pick by fit, not by loyalty.
| TVmaze API - TV Show and Episode Database | Internet Archive TV News Archive | |
|---|---|---|
| Publisher | TVmaze | Internet Archive |
| Subject lens | Programme catalogue: every TV show with seasons, episodes, cast, crew, AKAs, images, ratings and schedules | Archived television news: US national and local plus international broadcasters, searchable through closed-caption transcripts |
| Unit of analysis | Show, cascading to season, episode, person and credit level | One broadcast item - typically a half-hour to hour-long programme segment |
| Documented fields | 27 verified | 6 verified |
| Datadory rubric | 10/10 - the only perfect score among the 15 primary Broadcasting datasets | 9/10 |
| Best for | EPGs, trackers and catalogue apps needing machine-readable programme data and delta syncs | Fact-checking, media bias research and content analysis over real news footage and captions |
What each does better
TVmaze API - TV Show and Episode Database
A 27-field dictionary at programme grain. Every show record documents name, type (Scripted, Reality, Animation), language, genres, status (Running, Ended), runtime and averageRuntime, premiered and ended dates, officialSite and an HTML summary. Sample record Under the Dome: premiered June 24, 2013, ended September 10, 2015, rated 6.6, weight 99, Thursdays at 22:00 on CBS in America/New_York time. The news archive's dictionary runs six fields; nothing in it describes a programme this completely. Schedule intelligence no news archive attempts. schedule.time and schedule.days fix a show's usual slot, network.name carries an ISO country code and IANA timezone alongside webChannel for streaming carriers, and coverage extends to full historical episode archives plus future known air dates. For EPG-style builds that is EPG schedule data ready to model. Commercial and popularity signals. rating.average scores each show from user ratings and weight ranks relative importance - raw material for demand models and acquisition screens that caption text cannot supply. Cross-identifier resolution. externals.imdb, externals.thetvdb and externals.tvrage map every show onto the other big title keys, so a catalogue built here reconciles against film and episode databases without a fuzzy-match pass.
Internet Archive TV News Archive
Caption-level search over real airtime. Each item carries an identifier encoding channel and timestamp - sample record CNN_20010314_050000_Larry_King_Live, titled "Larry King Live : CNN : March 14, 2001 5:00am-6:00am EST" - plus channel, date, publicdate and description, all searchable through closed-caption text. That is what makes fact-checking, media-bias studies and phrase-tracking possible at the scale of millions of broadcasts; TVmaze holds a synopsis of a show but not one word spoken inside it, which is why closed-caption search has exactly one home in this pairing. A twenty-five-year news record. Holdings run from March 2001 through the present - about 4.37 million items as of August 2026 - beginning with US national networks plus San Francisco and Washington DC stations, expanded by roughly 40,000 tapes from Marion Stokes's estate covering Philadelphia and Boston, and now including international broadcasters such as M1 (Hungary), Press TV, Globovision, Syrian News, Zvezda and KQED. Recent samples reach August 21, 2026. Evidence rather than description. A broadcast item is the thing itself - programme, channel, air window, captions - so a claim like "this phrase aired on this channel on this date" resolves against a row, not against a plot summary.
Where they're equivalent
Same shelf, near-same grade. Both sit among Broadcasting's 15 primary datasets, both cleared field-definition verification during the same research pass in August 2026, and their rubric grades sit a single point apart - 10 versus 9 against the catalog's 7.81 average. That spine is why a title-channel-date match bridges the two, and why neither replaces the other past it. Structured rows with documented examples. Both dictionaries publish definitions and example values per field, so a sample cut settles within minutes whether the grain fits your model. Entity-grade identity. Version-stamped internal IDs and IMDb/TheTVDB/TVRage cross-references on one side; channel-plus-timestamp identifiers on the other. Either anchors a master list for its domain.
The verdict
Verdict: sample both, pick by fit - the question decides. Take TVmaze when the subject is the programme: building a viewing guide or tracker app, modelling genre and slot patterns from genres, schedule.time and schedule.days, watching status flip from Running to Ended, joining shows to people through cast and crew, benchmarking rating.average and weight for demand signals, or reconciling titles across IMDb, TheTVDB and TVRage keys. Take the TV News Archive when the subject is what was said: fact-checking a claim back to its airing, measuring how a frame or phrase travels across channels, studying media bias over a twenty-five-year span, tracking political advertising talk, or handing researchers clip-level evidence tied to channel and timestamp. Two quick tests settle most cases. Does your unit of analysis have episodes and cast? Only TVmaze describes those. Does it have spoken content? Only the Archive carries captions. Studying how entertainment schedules relate to what the news covered that night? That is the both-of-them case, and it is common.
Sample both, pick by fit. See TVmaze API - TV Show and Episode Database · See Internet Archive TV News Archive
Or take both in one feed
They stack as context and evidence rather than as a join. TVmaze fixes the grid - what airs where, on which network, in which country and timezone - then the TV News Archive supplies the text of what news programmes said inside that same broadcast day. A media-monitoring team can schedule around a show's slot and audit the newscasts adjacent to it; a political-science team can pair station-level news volume from the Archive with each market's entertainment landscape from TVmaze. Two practical notes from the records. There is no shared key between them, so the bridge is a deliberate title-channel-date match - workable because both sides name the programme and stamp the airtime, imperfect because TVmaze titles shows while the Archive titles segments. And mind the grain: hundreds of thousands of show records on one side, 4.37 million half-hour broadcasts on the other, so aggregate before comparing magnitudes. Or take both in one feed. Datadory normalizes each record to its documented dictionary - 27 fields on the programme side, 6 on the broadcast side - attaches sample rows for validation, and ships them beside the rest of the broadcasting catalog: delivered daily, weekly, or hourly, your call.
API, files, or your warehouse. Daily, weekly, or hourly.
Fair questions
Is TVmaze better than the Internet Archive TV News Archive for broadcasting analytics?
Different instruments. TVmaze wins whenever the question concerns programmes: hundreds of thousands of shows with seasons, episodes, cast, genres, user ratings and broadcast slots across a 27-field dictionary, scoring 10/10 on Datadory's rubric. The TV News Archive wins whenever the question concerns what news actually said: about 4.37 million broadcasts dated from March 2001 onward, searchable through closed-caption text, scoring 9/10.
Do TVmaze and the TV News Archive cover the same programmes?
Almost never. TVmaze catalogs entertainment and general programming - Scripted, Reality, Animation - with episode grids and cast lists. The TV News Archive holds television news programmes only, typically half-hour to hour-long segments from national networks and local stations. A nightly bulletin is the rare title both could describe, and even then one side carries its episode metadata while the other carries its caption text.
Which dataset reaches further back in time?
The TV News Archive, decisively. Its holdings run from March 2001 - early items include a CNN Larry King Live broadcast from March 14, 2001 - continuously to the present, with ingestion ongoing. TVmaze documents full historical episode archives for each show plus future air dates, but its depth depends on each title; there is no single year-zero comparable to the Archive's 2001 baseline.
Which has more documented fields, TVmaze or the TV News Archive?
TVmaze, by count and by depth: 27 verified fields covering titles, types, languages, genres, status, runtimes, premiere and end dates, schedule time and days, average rating, popularity weight, network and web channel, country and timezone, cross-IDs, images, synopsis and update stamps, against 6 verified fields describing identifier, title, channel, broadcast date, publicdate and description on each archived broadcast.
Can I search inside what was said on air with either dataset?
Only the Internet Archive TV News Archive supports that. Its records are searchable at caption-text level, which is what makes fact-checking and media-bias work possible at scale. TVmaze carries a synopsis of a show but no transcript of any moment inside it, so phrase-level questions about air content have no answer in its dictionary.
Can Datadory deliver both datasets together?
Yes. Sample both and pick by fit, or take both in one feed - normalized to their documented dictionaries (27 fields on the programme side, 6 on the broadcast side), aligned on whichever keys you choose, and validated against sample rows before anything ships.