Skip to main content

Importing Data

The Ingestion feature lets you import existing documents and web pages directly into Memsolus. Submit a file or a URL, and Memsolus extracts the content, splits it into meaningful chunks, and stores each chunk as an individual memory — ready to search immediately after processing.

Ingestion requires the ingest.write permission on your API key and the Ingest feature enabled on your plan.

Supported Formats​

FormatMax file size
PDF20 MB
DOCX20 MB
CSV10 MB
Markdown5 MB
JSON5 MB
Plain text (TXT)5 MB
HTML5 MB
URL (web page)—

How to Ingest a File​

1

Submit the file

Use a multipart/form-data POST request to submit your file along with the target user and optional categories.

curl -X POST https://api.memsolus.com/v1/memories/ingest \
  -H "X-Api-Key: msk_live_..." \
  -F "format=PDF" \
  -F "workspace_id=ws-uuid-here" \
  -F "user_id=user-123" \
  -F "categories=finance,reports" \
  -F "file=@/path/to/document.pdf"

You receive a job ID in the response:

{
  "id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "format": "PDF",
  "status": "PENDING",
  "user_id": "user-123",
  "categories": ["finance", "reports"],
  "memories_created": null,
  "created_at": "2026-01-15T10:30:00.000Z"
}
2

Poll for completion

The job is processed asynchronously. Poll the job status until it reaches COMPLETED or FAILED.

curl "https://api.memsolus.com/v1/ingest-jobs/a1b2c3d4-e5f6-7890-abcd-ef1234567890" \
  -H "X-Api-Key: msk_live_..."
{
  "id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "format": "PDF",
  "status": "COMPLETED",
  "memories_created": 34,
  "updated_at": "2026-01-15T10:32:00.000Z"
}
3

Search the ingested memories

Once status is COMPLETED, the created memories are searchable immediately.

curl -X POST https://api.memsolus.com/v1/memories/search \
  -H "X-Api-Key: msk_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "query": "quarterly revenue",
    "user_id": "user-123",
    "categories": ["finance"]
  }'

How to Ingest a URL​

Submit a public URL instead of a file. Memsolus fetches the page, extracts the text, and creates memories from the content.

curl -X POST https://api.memsolus.com/v1/memories/ingest \
  -H "X-Api-Key: msk_live_..." \
  -H "Content-Type: multipart/form-data" \
  -F "format=URL" \
  -F "workspace_id=ws-uuid-here" \
  -F "user_id=user-123" \
  -F "source_url=https://example.com/article"

Using the SDK​

import { Memsolus } from '@memsolus/sdk';
import { readFileSync } from 'fs';

const client = new Memsolus({ apiKey: 'msk_live_...' });

const job = await client.ingest.create({
  format: 'PDF',
  workspaceId: 'ws-uuid-here',
  userId: 'user-123',
  file: readFileSync('/path/to/document.pdf'),
  categories: ['finance', 'reports'],
});

// Poll until done
const result = await client.ingest.waitForCompletion(job.id);
console.log(`Created ${result.memoriesCreated} memories`);

Job Status​

StatusWhat it means
PENDINGJob received, waiting to start
PROCESSINGExtracting and chunking content
COMPLETEDAll memories created and searchable
FAILEDCould not process the file — check the error field for details

A typical 20-page PDF completes in under 30 seconds.

Limits and Notes​

  • All memories created by an ingest job share the same user_id and categories you specify
  • If the file is password-protected, encrypted, or corrupt, the job fails with a descriptive error message