scriptoriumadmin/supermemory
1
0
Fork 0
mirror of https://github.com/supermemoryai/supermemory.git synced 2026-09-08 22:21:07 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Delete apps/docs/search/overview.mdx

Browse source 
Co-Authored-By: mintlify[bot] <109931778+mintlify[bot]@users.noreply.github.com>
This commit is contained in:
mintlify[bot] 2025-12-19 00:31:40 +00:00 • committed by GitHub
parent 3dfb3e7823
commit 856e06dfc8
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 0 additions and 87 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

87
apps/docs/search/overview.mdx
View file

@ -1,87 +0,0 @@
---
title: Search memories
description: Search through your stored memories and documents using semantic search
---
The v4 search API allows you to search through your stored memories and documents using semantic search. It supports two search modes: memories-only search and hybrid search.
## Search modes
### Memories mode (default)
Searches only through memory entries that have been stored in your organization. This is the default mode and is optimized for finding specific memories.
```json
{
"q": "What are John's preferences?",
"searchMode": "memories"
}
```
### Hybrid mode
Hybrid search provides a fallback mechanism when memories don't contain the information you're looking for. It works in two stages:
1. **Memory search**: First searches through your stored memories
2. **Chunk fallback**: If no or insufficient memories are found, searches through document chunks
This mode is useful when you want comprehensive results that include both structured memories and raw document content.
```json
{
"q": "What are John's preferences?",
"searchMode": "hybrid"
}
```
## How hybrid search works
When using hybrid search mode:
1. The system performs a semantic search across your memory entries
2. In parallel, it searches through document chunks using vector similarity
3. Results are merged and sorted by similarity score
4. Duplicate chunks (those already associated with memory results) are automatically removed
5. The top results are returned, respecting your specified limit
This approach ensures you get the most relevant information whether it's stored as a structured memory or within a document chunk.
## Key features
- **Semantic search**: Uses embeddings to find semantically similar content
- **Query rewriting**: Optionally rewrites queries for better search results
- **Reranking**: Optionally reranks results for improved relevance
- **Filtering**: Filter by metadata, date ranges, and container tags
- **Context inclusion**: Include related memories, documents, and chunks
- **Temporal context**: Include memories created before/after results
## Quick example
```bash
curl -X POST https://api.supermemory.ai/v4/search \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"q": "machine learning preferences",
"searchMode": "hybrid",
"limit": 10,
"threshold": 0.7
}'
```
## Next steps
<CardGroup cols={2}>
<Card title="Parameters" icon="sliders" href="/search/parameters">
Learn about all available search parameters
</Card>
<Card title="Response schema" icon="brackets-curly" href="/search/response-schema">
Understand the response structure
</Card>
<Card title="Examples" icon="code" href="/search/examples/memory-search">
See practical examples
</Card>
<Card title="Filtering" icon="filter" href="/search/filtering">
Learn about advanced filtering
</Card>
</CardGroup>