Trusted by leading media companies
Legacy search wasn't built for media
Huge archives, mixed formats, breaking news, and audiences who search the way they speak.
Video on demand
500k monthly users, one index
One of the largest French news and sports broadcasters runs search, recommendations and navigation for its video platform from a single Meilisearch index.
Champions League highlights
Champions League: goals of the week
Champions League draw explained
Recommended for you
Four intents. One engine.
Search
They know what they want, even when they misspell it.
- Typo-tolerant full-text search
- Results as you type
- 100+ languages
Discovery
They describe it instead of naming it.
- Semantic search by description
- Hybrid full-text and semantic
- Similar content from any item
Recommendation
They didn't ask, but want a personal experience.
- Personalized reranking
- Pinned results editors manage
- Ranking by freshness or popularity
Chat
A short conversation beats a long search.
- Conversational search API
- MCP server for AI agents
- Answers grounded in your catalog
Two audiences, one index
Viewer apps and newsroom tools read from the same index. Here is how your content flows through it.
1Sources
2Index
3Surfaces
Video and audio
Clips, rushes, segments
Text and transcripts
Articles, scripts, subtitles
Asset metadata
MAM, CMS: who, when, rights
Meilisearch index
- Hybrid full-text + semantic
- Vectors from any model
- Filters and facets on metadata
- Real-time indexing
Viewer apps
Search, discovery, rails
Newsroom and archive
Find past coverage fast
Production and post
Find footage by description
AI agents and chat
Grounded answers via MCP
Enrichment vendors
Transcription, OCR, visual embeddings
1Sources
Video and audio
Clips, rushes, segments
Text and transcripts
Articles, scripts, subtitles
Asset metadata
MAM, CMS: who, when, rights
Enrichment vendors
Transcription, OCR, visual embeddings
Meilisearch index
- Hybrid full-text + semantic
- Vectors from any model
- Filters and facets on metadata
- Real-time indexing
3Surfaces
Viewer apps
Search, discovery, rails
Newsroom and archive
Find past coverage fast
Production and post
Find footage by description
AI agents and chat
Grounded answers via MCP
Pick a step to see how content flows through one index.
Typical responses under 50ms · Up to 99.999% uptime on Cloud
Integration · TwelveLabs
Search what's in the video, not just its title
Index embeddings from video models such as TwelveLabs Marengo next to your metadata. No tags yet? Transcripts, OCR and visual embeddings become searchable too.
Read the TwelveLabs guide →- Describe the scene
- Search by example
- Find the moment
Plug into your AI stack
Agents, chat and RAG pipelines get the same retrieval your audience uses.
Agents query your catalog
A built-in MCP server exposes your indexes as tools, so AI agents search, filter and cite your content directly.
{
"indexUid": "articles",
"q": "rail strike",
"filter": "section = 'news'",
"limit": 3
}3 articles found. Talks resumed on day two, with services cut on most lines.
On-demand talk
Search, discovery and recommendations from one index
The talk Quentin de Quelen, CEO and co-founder of Meilisearch, gave at IBC 2026: a demo in production on real media catalogs.
What you'll see
- The four intents your audience brings, and what happens when you miss them
- How RMC+ serves search, recommendations and navigation to 500k monthly users from one index
- A demo switching between full-text and semantic search on a movie catalog
FAQ