Navigation

Mobile Docs

RAG Agent

Retrieval-Augmented Generation agent

Create a RAG (Retrieval-Augmented Generation) agent that retrieves information from document sources to provide accurate, grounded responses.

Agentsrag_agent

Usage

01Add the block to your workflow and connect it to the upstream step.
02Configure any required credentials or tokens in the inputs.
03Fill in required inputs and optional parameters for the run.
04Run a test execution, inspect outputs, and iterate before deploying.
05Deploy the rag_agent block with monitoring enabled in production.

Inputs (UI)

Agent Profile

agent-profile-selector

Layout: full

System Prompt

long-input

Placeholder: Enter system prompt for the RAG agent...

Layout: full

Condition: agentProfileId = ""

Query

long-input

Placeholder: Enter the query to search for information...

Layout: full

Model

dropdown

Layout: full

Options: Dynamic options

Data Source

dropdown

Layout: half

Options: Pinecone, Supabase, MongoDB, Airtable

Data Source Configuration

code

Placeholder: { "connectionString": "your-connection-string", "collection": "your-collection-name" }

Layout: full

Retrieval Options

code

Placeholder: { "topK": 5, "similarityThreshold": 0.7, "includeMetadata": true }

Layout: full

Cite Sources

switch

Layout: half

Enable Memory

switch

Layout: half

Memory Limit

slider

Layout: half

Condition: memoryEnabled = true

Importance Threshold

slider

Layout: half

Condition: memoryEnabled = true

Memory Tags

short-input

Placeholder: customer-support, billing, technical

Layout: half

Condition: memoryEnabled = true

Inputs (API)

agentProfileId

string

Optional

systemPrompt

string

Optional

query

string

Required

model

string

Required

apiKey

string

Optional

dataSource

string

Required

dataSourceConfig

json

Required

retrievalOptions

json

Optional

temperature

number

Optional

citeSources

boolean

Optional

memoryEnabled

boolean

Optional

memoryLimit

number

Optional

memoryImportance

number

Optional

memoryTags

string

Optional

Outputs

Primary response type:

{
  "content": "string",
  "model": "string",
  "sources": "json",
  "tokens": "json"
}

Tool Access

openai_chatanthropic_chatgoogle_chatxai_chatdeepseek_chatpineconesupabasemongodbairtable