AmaStudio Knowledge Sync

AmaStudio can sync knowledge bases with amadev.org, enabling vector search across your documents.

How It Works

Create a Knowledge Base — A container for related documents
Upload Documents — PDF, text, or markdown files
Automatic Chunking — Documents are split into chunks
Embedding — Each chunk is converted to a vector embedding
Search — Query the knowledge base using semantic search

API Endpoints

Create a Knowledge Base

Code

POST /v1/knowledge
Content-Type: application/json

{
  "name": "Project Docs",
  "description": "Internal project documentation",
  "embeddingModel": "openai/text-embedding-3-small"
}

List Knowledge Bases

Code

GET /v1/knowledge

Get Knowledge Base Details

Code

GET /v1/knowledge/:id

Upload a Document

Code

POST /v1/knowledge/:id/documents
Content-Type: multipart/form-data

file: @document.pdf

Search a Knowledge Base

Code

POST /v1/knowledge/:id/search
Content-Type: application/json

{
  "query": "deployment instructions",
  "limit": 10
}

Delete a Knowledge Base

Code

DELETE /v1/knowledge/:id

Embedding Models

AmaStudio supports multiple embedding providers:

Provider	Model	Use Case
OpenAI	`text-embedding-3-small`	Production (default)
VoyageAI	`voyage-2`	High-quality embeddings
Ollama	`nomic-embed-text`	Local development

The embedding provider is configured per knowledge base. If no API key is specified for the KB, the server falls back to OPENAI_API_KEY from environment variables.

Sync from AmaStudio Desktop

AmaStudio's AmadevService provides IPC methods for knowledge sync:

AmadevSyncKnowledge — triggers a full sync
Knowledge bases, documents, and chunks are synced to amadev.org
Synced data is searchable from any AmaStudio instance