Documentation Search
Qarion indexes its own documentation — user guides and developer API references — into Elasticsearch for instant full-text and semantic search. This allows users and the AI Copilot to find relevant documentation quickly.
Searching Documentation
Documentation search is publicly accessible — no authentication required. You can search via:
API
curl "https://api.qarion.com/docs/search?q=quality+checks&mode=auto&limit=10"
Copilot
The AI Copilot automatically uses documentation search as one of its tools. When you ask the Copilot a question about how to use a feature, it can search the docs and cite relevant sections in its response.
Search Modes
| Mode | Description |
|---|---|
| Auto | (default) Combines keyword and semantic results |
| Keyword | BM25 full-text search in Elasticsearch |
| Semantic | Vector-similarity search using AI embeddings |
Filtering by Source
You can filter results to specific documentation sources:
| Source | Description |
|---|---|
all | Search across all documentation (default) |
user | User guides only (docs.qarion.com) |
developer | Developer API references only (developers.qarion.com) |
Reindexing
When documentation content is updated, the index needs to be refreshed. Administrators can trigger a reindex from the admin panel or via the API:
curl -X POST "https://api.qarion.com/docs/reindex" \
-H "Authorization: Bearer YOUR_API_KEY"
The reindex operation:
- Scans markdown files from both documentation sites
- Splits content into searchable chunks (by section)
- Generates vector embeddings for semantic search
- Indexes everything into Elasticsearch
Related
- Global Search — Platform-wide entity search
- AI Copilot — How the Copilot uses documentation search