> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bkmark.it/llms.txt
> Use this file to discover all available pages before exploring further.
# Search Your Bookmark Library: Full-Text and Semantic
> Find anything instantly with full-text search on all plans, or upgrade to Pro for AI-powered semantic search that understands meaning, not just keywords.
Bkmark's search is built to find the right bookmark fast — whether you remember the exact title, a word from the description, or just the general idea of what you're looking for. Every plan includes full-text search, and Pro and Team users get an additional layer of AI-powered semantic understanding.
***
## How search works
### Full-text search (all plans)
Full-text search scans the title, description, and URL of every bookmark in your library. It is fast, always available, and works on every plan.
* **Exact and partial matches** — results include bookmarks where your query terms appear as whole words or as stemmed variants. Searching `typescript` will also surface `TypeScript` and related forms.
* **Fuzzy fallback** — if there are no exact matches, Bkmark automatically falls back to fuzzy matching so you still get relevant results even with typos or alternate spellings.
* **Fast** — results appear in milliseconds regardless of how large your library is.
### AI semantic search (Pro & Team)
AI semantic search requires a **Pro** or **Team** plan. On the Free plan, you get full-text search only.
Semantic search understands the *meaning* behind your query rather than just matching characters. You can search using natural language descriptions, and Bkmark will surface bookmarks that are conceptually related — even if they don't share a single word with your query.
**Example:** Searching for `"how to speed up database queries"` with semantic search can surface bookmarks about query optimization, indexing strategies, and caching — even if those pages never use the phrase "speed up."
***
## Exact search vs. semantic search — what's the difference?
| | Full-text (all plans) | Semantic search (Pro/Team) |
| ------------------ | ----------------------------- | ----------------------------------------------- |
| **How it works** | Matches words and their stems | Matches meaning and concepts |
| **Best for** | Known terms, titles, URLs | Vague recollections, concepts, natural language |
| **Typo tolerance** | Yes (fuzzy fallback) | Yes (intent is preserved) |
| **Speed** | Very fast (milliseconds) | Fast (AI-assisted) |
| **Requires** | Free plan or above | Pro or Team plan |
***
## Using search
### Search from the app
Click the **Search** bar at the top of the sidebar (or press `/` from anywhere in the app) and start typing. Results update live as you type.
### Search filters
Narrow your results using the filter controls next to the search bar:
| Filter | Description |
| --------- | --------------------------------------------------- |
| **Group** | Show only results inside a specific group |
| **Tag** | Show only results with a specific tag |
| **Type** | Limit results to `link`, `note`, `code`, or `image` |
You can combine filters with your query — for example, search for `"authentication"` filtered to the `Engineering` group and the `security` tag.
### Using the search API
You can query your library programmatically using the search endpoint.
```bash theme={null}
curl "https://api.bkmark.it/api/v1/search?q=typescript+patterns" \
-H "Authorization: Bearer YOUR_ACCESS_TOKEN"
```
**Query parameters:**
| Parameter | Type | Default | Description |
| --------- | ------ | -------- | ------------------------------------------ |
| `q` | string | Required | Your search query |
| `cursor` | string | — | Pagination cursor from a previous response |
| `limit` | number | `20` | Results per page (1–100) |
| `groupId` | string | — | Filter by group UUID |
| `tagId` | string | — | Filter by tag UUID |
**Example response:**
```json theme={null}
{
"data": [
{
"id": "3f7a9c12-...",
"url": "https://www.typescriptlang.org/docs/handbook/2/types-from-types.html",
"title": "TypeScript Handbook: Creating Types from Types",
"description": "A deep dive into mapped types, conditional types, and template literal types.",
"type": "link",
"isFavorite": false,
"isArchived": false,
"tags": [{ "id": "uuid", "name": "typescript", "color": "blue" }],
"groups": [{ "id": "uuid", "name": "Engineering References" }],
"createdAt": "2025-03-15T10:22:00.000Z",
"updatedAt": "2025-03-15T10:22:00.000Z"
}
],
"nextCursor": "3f7a9c12-..."
}
```
### Paginating results
Search uses cursor-based pagination. When a response includes a `nextCursor` value, pass it as the `cursor` parameter in your next request to fetch the following page. When `nextCursor` is `null`, you've reached the last page.
```bash theme={null}
# Page 1
curl "https://api.bkmark.it/api/v1/search?q=design+systems&limit=20" \
-H "Authorization: Bearer YOUR_ACCESS_TOKEN"
# Page 2 — pass nextCursor from the previous response
curl "https://api.bkmark.it/api/v1/search?q=design+systems&limit=20&cursor=3f7a9c12-..." \
-H "Authorization: Bearer YOUR_ACCESS_TOKEN"
```
***
## Tips for effective searching
On Pro or Team, try natural-language queries like `"articles about building design systems in React"` instead of guessing the exact title you used when saving.
Pair a short query with a group or tag filter to dramatically narrow results — especially useful when you have a large library with overlapping topics.
On full-text search, two or three specific words from the title or description work better than long sentences. Try `"react server components"` instead of `"how to use react server components in next"`.
Remember the site but not the article? Search for the domain name (e.g., `"vercel.com"` or `"css-tricks"`) to surface all bookmarks from that source.