mirror of
https://github.com/supermemoryai/supermemory.git
synced 2026-08-28 05:25:33 +00:00
59 lines
No EOL
1.1 KiB
Text
59 lines
No EOL
1.1 KiB
Text
---
|
|
openapi: post /add
|
|
---
|
|
|
|
Add a new memory with content and metadata.
|
|
|
|
Fields:
|
|
|
|
`content`: string
|
|
|
|
`id`: string
|
|
|
|
`metadata`: Record
|
|
|
|
The `content` can be of the following types:
|
|
|
|
- note \/ Markdown
|
|
|
|
- If it is a markdown, all the images inside `![]` image tags will automatically be parsed.
|
|
|
|
- pdf
|
|
|
|
- tweet
|
|
|
|
- google_doc
|
|
|
|
- notion_doc
|
|
|
|
- webpage URL
|
|
|
|
- Images and other content is also intelligently parsed in case of a webpage.
|
|
|
|
|
|
The metadata provided is a JSON object.
|
|
|
|
for eg.
|
|
|
|
``` json
|
|
{
|
|
"classId": "21412",
|
|
"year": "fifth"
|
|
}
|
|
|
|
```
|
|
|
|
If you wish to do exact searches, please use strings. But if you want to search in a range (time, numbers, prices), you can use numbers too.
|
|
|
|
``` json
|
|
{
|
|
"price": 1250
|
|
}
|
|
|
|
```
|
|
|
|
More about \[metadata filtering here\]([https://docs.supermemory.ai/essentials/metadata-filtering](https://docs.supermemory.ai/essentials/metadata-filtering))
|
|
|
|
The `id` is optional. If provided, supermemory will store the same ID as your internal database. This can help for retrieval purposes.
|
|
|
|
If the `id` already exists, supermemory will update it instead. |