Knowledge

Build knowledge bases with vector storage, embeddings, and content ingestion for retrieval-augmented generation (RAG).

Knowledge in Agno is composed of several resources that work together:

Content (URLs, text)
  └─▶ Knowledge
        ├─▶ VectorDB (Qdrant)
        │     └─▶ Embedder (OpenAI)
        └─▶ Contents DB (Postgres, optional)

`agno/knowledge`

The central knowledge base resource that connects a vector database with optional embedder and content storage.

Config

Field	Type	Required	Default	Description
`vector_db`	`Dependency`	Yes	—	Reference to `agno/vectordb/qdrant`
`contents_db`	`Dependency`	No	—	Reference to `agno/db/postgres` for content metadata storage
`embedder`	`Dependency`	No	—	Reference to `agno/knowledge/embedder/openai` for custom embeddings
`max_results`	`integer`	No	`10`	Maximum search results to return

Outputs

Field	Type	Description
`pip_dependencies`	`list[string]`	Required Python packages
`spec`	`object`	Serialized knowledge spec for runtime reconstruction

`agno/vectordb/qdrant`

Configures a Qdrant vector database for storing embeddings.

Config

Field	Type	Required	Default	Description
`url`	`Field[string]`	Yes	—	Qdrant server URL. Supports FieldReferences
`collection`	`Field[string]`	Yes	—	Collection name. Supports FieldReferences
`api_key`	`Field[string]`	No	—	API key for authentication
`search_type`	`"vector"` \| `"keyword"` \| `"hybrid"`	No	`"hybrid"`	Search strategy
`embedder`	`Dependency`	No	—	Reference to `agno/knowledge/embedder/openai`

Outputs

Field	Type	Description
`spec`	`object`	Serialized Qdrant config for runtime reconstruction
`pip_dependencies`	`list[string]`	Required packages (includes `fastembed` for hybrid/keyword search)

`agno/knowledge/embedder/openai`

Configures an OpenAI embedding model for generating vector embeddings.

Config

Field	Type	Required	Default	Description
`id`	`string`	No	`"text-embedding-3-small"`	Embedding model identifier
`api_key`	`Field[string]`	Yes	—	OpenAI API key
`dimensions`	`integer`	No	—	Override embedding dimensions (only for `text-embedding-3-*` models)
`encoding_format`	`"float"` \| `"base64"`	No	`"float"`	Response encoding format
`organization`	`string`	No	—	OpenAI organization ID
`base_url`	`string`	No	—	Custom base URL for OpenAI-compatible APIs

Outputs

Field	Type	Description
`pip_dependencies`	`list[string]`	Required Python packages
`spec`	`object`	Serialized embedder spec for runtime reconstruction

`agno/knowledge/content`

Represents a content source to ingest into a knowledge base. Supports URLs and inline text.

Config

Field	Type	Required	Default	Description
`knowledge`	`Dependency`	Yes	—	Reference to `agno/knowledge`
`url`	`string`	Conditional	—	URL to fetch (website, PDF, remote file). Mutually exclusive with `text_content`
`text_content`	`string`	Conditional	—	Inline text content. Mutually exclusive with `url`
`name`	`string`	No	Resource name	Content identifier
`description`	`string`	No	—	Content description
`metadata`	`dict[string, string]`	No	—	Custom metadata key-value pairs
`topics`	`list[string]`	No	—	Topic tags for categorization

Outputs

Field	Type	Description
`spec`	`object`	Content specification
`pip_dependencies`	`list[string]`	Reader packages (`pypdf`, `beautifulsoup4`, `python-docx`, etc.)

Example

A complete knowledge base setup:

# 1. Embedder
provider: agno
resource: knowledge/embedder/openai
name: embedder
config:
  id: text-embedding-3-small
  api_key:
    provider: pragma
    resource: secret
    name: openai-key
    field: outputs.OPENAI_API_KEY
---
# 2. Vector database
provider: agno
resource: vectordb/qdrant
name: vectors
config:
  url:
    provider: qdrant
    resource: database
    name: main
    field: outputs.url
  collection:
    provider: qdrant
    resource: collection
    name: docs
    field: outputs.name
  api_key:
    provider: qdrant
    resource: database
    name: main
    field: outputs.api_key
  search_type: hybrid
  embedder:
    provider: agno
    resource: knowledge/embedder/openai
    name: embedder
---
# 3. Knowledge base
provider: agno
resource: knowledge
name: docs-kb
config:
  vector_db:
    provider: agno
    resource: vectordb/qdrant
    name: vectors
  max_results: 5
---
# 4. Content source
provider: agno
resource: knowledge/content
name: user-guide
config:
  knowledge:
    provider: agno
    resource: knowledge
    name: docs-kb
  url: "https://docs.example.com/guide.pdf"
  topics:
    - documentation
    - user-guide

Notes

You must provide exactly one of url or text_content in content resources.
Content supports URLs to PDFs, web pages, documents (DOCX, PPTX), arxiv papers, YouTube transcripts, and Wikipedia articles.
The search_type: hybrid option combines vector and keyword search for better results but requires the fastembed package.
Content resources are stateful — they insert into and delete from the vector database during lifecycle events.

Getting Started

Concepts

Resource Catalog

Agno (AI Agents)

GCP

Kubernetes

Building Providers

`agno/knowledge`

Config

Outputs

`agno/vectordb/qdrant`

Config

Outputs

`agno/knowledge/embedder/openai`

Config

Outputs

`agno/knowledge/content`

Config

Outputs

Example

Notes

​agno/knowledge

​Config

​Outputs

​agno/vectordb/qdrant

​Config

​Outputs

​agno/knowledge/embedder/openai

​Config

​Outputs

​agno/knowledge/content

​Config

​Outputs

​Example

​Notes

`agno/knowledge`

Config

Outputs

`agno/vectordb/qdrant`

Config

Outputs

`agno/knowledge/embedder/openai`

Config

Outputs

`agno/knowledge/content`

Config

Outputs

Example

Notes