Workflows

Build and deploy automated pipelines with the Case.dev Visual Workflow Builder. Workflows run on Relay—our serverless execution engine that scales from zero to millions of executions.

Overview

Workflows chain Case.dev services (Vaults, LLMs, OCR, Search, Voice, Format) into automated pipelines:

Visual Builder — Drag and drop nodes at workflow.case.dev
Webhook Triggers — Every workflow gets a unique webhook URL
Serverless Execution — Pay only for what you use, automatic scaling
Full Observability — Monitor every execution with detailed logs

Quick Start

Option 1: Visual Builder (Recommended)

Open workflow.case.dev
Drag action nodes onto the canvas
Connect nodes with edges
Configure each node’s settings
Click Deploy to get your webhook URL

Option 2: API-First (Programmatic)

Create and deploy a workflow in a single API call:

curl -X POST https://api.case.dev/workflows/v1/create \
  -H "Authorization: Bearer $CASEDEV_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Document Summary Pipeline",
    "description": "Search and summarize documents",
    "nodes": [
      {
        "id": "trigger-1",
        "type": "trigger",
        "label": "Webhook Trigger",
        "config": { "triggerType": "Webhook" }
      },
      {
        "id": "search-1",
        "type": "action",
        "label": "Search Web",
        "config": {
          "actionType": "case-search/web-search",
          "query": "{{topic}}"
        }
      },
      {
        "id": "llm-1",
        "type": "action",
        "label": "Summarize",
        "config": {
          "actionType": "case-llm/llm",
          "method": "chat",
          "model": "gpt-4o",
          "systemPrompt": "Summarize the search results concisely.",
          "userPrompt": "{{results.Search_Web.output.results}}"
        }
      }
    ],
    "edges": [
      { "source": "trigger-1", "target": "search-1" },
      { "source": "search-1", "target": "llm-1" }
    ]
  }'

Response:

{
  "id": "wf_abc123",
  "name": "Document Summary Pipeline",
  "webhookUrl": "https://api.case.dev/workflows/v1/wf_abc123/webhook",
  "webhookSecret": "whsec_...",
  "status": "deployed"
}

Important: Save your webhookSecret—it won’t be shown again.

Execute Your Workflow

curl -X POST "https://api.case.dev/workflows/v1/wf_abc123/webhook" \
  -H "X-Webhook-Secret: whsec_..." \
  -H "Content-Type: application/json" \
  -d '{"topic": "AI in legal tech 2025"}'

Template Syntax

Reference data from the trigger input and previous steps using {{template}} syntax:

Trigger Input

Access fields from the webhook request body directly:

{{fieldName}}       → Value from webhook body
{{nested.field}}    → Nested fields work too

Previous Step Output

Access results from earlier steps in the workflow:

{{results.StepLabel.output}}              → Entire output object
{{results.StepLabel.output.fieldName}}    → Specific field
{{results.StepLabel.output.items[0]}}     → Array access

Examples:

{{results.Search_Web.output.results}}
{{results.Summarize.output.choices[0].message.content}}
{{results.OCR_Document.output.text}}

Note: Step labels with spaces become underscores: “Search Web” → Search_Web

Available Action Types

case-llm/llm

Generate text with LLMs (OpenAI, Anthropic, etc.)

Field	Description
`method`	`chat`, `complete`, or `list-models`
`model`	Model ID (`gpt-4o`, `claude-3-5-sonnet`, etc.)
`systemPrompt`	System instructions
`userPrompt`	User message with templates
`temperature`	0-2, controls randomness
`maxTokens`	Maximum output tokens

case-search/web-search

Search the web for real-time information.

Field	Description
`query`	Search query
`searchType`	`auto`, `general`, `news`, or `academic`
`numResults`	Number of results (default: 5)

case-vault/search

Search documents in a vault.

Field	Description
`vaultId`	Vault ID
`query`	Search query
`method`	`hybrid`, `vector`, `graph`, or `entity`
`limit`	Max results (default: 10)

case-ocr/process

Extract text from images or scanned PDFs.

Field	Description
`documentUrl`	URL to document
`engine`	`gpt-4o`, `doctr`, or `paddleocr`

case-voice/transcribe

Transcribe audio files.

Field	Description
`audioUrl`	URL to audio file
`model`	`whisper-1` or `whisper-large-v3`
`language`	Optional language code

case-format/format

Convert content to PDF, DOCX, or HTML.

Field	Description
`content`	Markdown or text content
`inputFormat`	`md` or `txt`
`outputFormat`	`pdf`, `docx`, or `html`

HTTP Request

Make custom HTTP requests to external APIs.

Field	Description
`method`	`GET`, `POST`, `PUT`, `PATCH`, `DELETE`
`url`	Target URL
`body`	Request body (JSON)
`headers`	Custom headers (JSON)

Execution Modes

Fire-and-Forget (Default)

Trigger the workflow and check status later:

curl -X POST ".../webhook" \
  -H "X-Webhook-Secret: whsec_..." \
  -d '{"input": "data"}'

Response includes executionId for status checks.

Synchronous (Wait for Result)

Block until the workflow completes:

curl -X POST ".../webhook?mode=sync" \
  -H "X-Webhook-Secret: whsec_..." \
  -d '{"input": "data"}'

Returns the final output directly.

Callback Webhook

Get notified when complete:

curl -X POST ".../webhook" \
  -H "X-Webhook-Secret: whsec_..." \
  -H "X-Callback-Url: https://your-app.com/webhook" \
  -d '{"input": "data"}'

Your callback URL receives the final result.

Example Workflows

Document Analysis Pipeline

{
  "name": "Document Analysis",
  "nodes": [
    {
      "id": "trigger",
      "type": "trigger",
      "label": "Webhook",
      "config": { "triggerType": "Webhook" }
    },
    {
      "id": "ocr",
      "type": "action",
      "label": "Extract Text",
      "config": {
        "actionType": "case-ocr/process",
        "documentUrl": "{{documentUrl}}",
        "engine": "gpt-4o"
      }
    },
    {
      "id": "analyze",
      "type": "action",
      "label": "Analyze",
      "config": {
        "actionType": "case-llm/llm",
        "method": "chat",
        "model": "gpt-4o",
        "systemPrompt": "You are a document analyst. Extract key facts, dates, and parties.",
        "userPrompt": "Analyze this document:\n\n{{results.Extract_Text.output.text}}"
      }
    }
  ],
  "edges": [
    { "source": "trigger", "target": "ocr" },
    { "source": "ocr", "target": "analyze" }
  ]
}

Research Agent

{
  "name": "Research Agent",
  "nodes": [
    {
      "id": "trigger",
      "type": "trigger",
      "label": "Webhook",
      "config": { "triggerType": "Webhook" }
    },
    {
      "id": "search",
      "type": "action",
      "label": "Research",
      "config": {
        "actionType": "case-search/web-search",
        "query": "{{query}}",
        "numResults": "10"
      }
    },
    {
      "id": "synthesize",
      "type": "action",
      "label": "Synthesize",
      "config": {
        "actionType": "case-llm/llm",
        "method": "chat",
        "model": "gpt-4o",
        "systemPrompt": "Synthesize the research into a comprehensive summary with citations.",
        "userPrompt": "Research results:\n{{results.Research.output.results}}"
      }
    },
    {
      "id": "format",
      "type": "action",
      "label": "Create PDF",
      "config": {
        "actionType": "case-format/format",
        "content": "# Research Report\n\n{{results.Synthesize.output.choices[0].message.content}}",
        "inputFormat": "md",
        "outputFormat": "pdf"
      }
    }
  ],
  "edges": [
    { "source": "trigger", "target": "search" },
    { "source": "search", "target": "synthesize" },
    { "source": "synthesize", "target": "format" }
  ]
}

Observability

Monitor your workflow executions in the Visual Builder at workflow.case.dev/runs:

Execution History — See all runs with status, duration, timestamps
Step Details — Inspect input/output for each step
Error Logs — Debug failed executions

Or query via API:

# List executions for a workflow
GET /workflows/v1/{id}/runs

# Get execution details
GET /workflows/v1/runs/{executionId}

Pricing

Workflows are billed per execution based on underlying service usage:

Relay Execution — $0.0001 per step execution
LLM Tokens — Standard token pricing
OCR Pages — Standard page pricing
Search Queries — Standard query pricing

View detailed usage in the Console.

Next Steps

API Reference — Full endpoint documentation
Open Visual Builder — Start building
Cookbooks — Example workflows for common use cases

Workflows can orchestrate any Case.dev service:

Vaults

Search and retrieve documents from your case vaults

OCR

Extract text from scanned documents and images

Voice

Transcribe audio files with speaker diarization

LLMs

Generate, analyze, and extract with language models

Web Search

Search the web and get AI-powered answers with citations

Format

Generate PDFs, Word docs, and HTML from workflow outputs

Get Started

Platform

Resources

Workflows

Workflows

Overview

Quick Start

Option 1: Visual Builder (Recommended)

Option 2: API-First (Programmatic)

Execute Your Workflow

Template Syntax

Trigger Input

Previous Step Output

Available Action Types

case-llm/llm

case-search/web-search

case-vault/search

case-ocr/process

case-voice/transcribe

case-format/format

HTTP Request

Execution Modes

Fire-and-Forget (Default)

Synchronous (Wait for Result)

Callback Webhook

Example Workflows

Document Analysis Pipeline

Research Agent

Observability

Pricing

Next Steps

Vaults

OCR

Voice

LLMs

Web Search

Format

Get Started

Platform

Resources

​Workflows

​Overview

​Quick Start

​Option 1: Visual Builder (Recommended)

​Option 2: API-First (Programmatic)

​Execute Your Workflow

​Template Syntax

​Trigger Input

​Previous Step Output

​Available Action Types

​case-llm/llm

​case-search/web-search

​case-vault/search

​case-ocr/process

​case-voice/transcribe

​case-format/format

​HTTP Request

​Execution Modes

​Fire-and-Forget (Default)

​Synchronous (Wait for Result)

​Callback Webhook

​Example Workflows

​Document Analysis Pipeline

​Research Agent

​Observability

​Pricing

​Next Steps

​Related services

Vaults

OCR

Voice

LLMs

Web Search

Format

Workflows

Overview

Quick Start

Option 1: Visual Builder (Recommended)

Option 2: API-First (Programmatic)

Execute Your Workflow

Template Syntax

Trigger Input

Previous Step Output

Available Action Types

case-llm/llm

case-search/web-search

case-vault/search

case-ocr/process

case-voice/transcribe

case-format/format

HTTP Request

Execution Modes

Fire-and-Forget (Default)

Synchronous (Wait for Result)

Callback Webhook

Example Workflows

Document Analysis Pipeline

Research Agent

Observability

Pricing

Next Steps

Related services