digital-labour（数字化劳动力）

⚡ Digital Labour — 24 AI Agents for Business Automation

Your entire back-office, automated. Run any of 24 specialized AI agents through plain English. Sales outreach, lead generation, content repurposing, SEO, ad copy, bookkeeping, proposals, market research, tech docs — all with built-in QA verification.

Quick Start

Set your API URL (or use the default production endpoint):

export DIGITAL_LABOUR_API_URL="https://bitrage-labour-api-production.up.railway.app"

Test the connection:

python3 {baseDir}/scripts/dl-api.py health

Run an agent:

python3 {baseDir}/scripts/dl-api.py run support_ticket '{"ticket_text": "My order has not arrived after 2 weeks"}'

Available Agents (24 Total)

Revenue & Sales (5 agents)

Content & Marketing (4 agents)

Operations & Data (4 agents)

Documents & Research (4 agents)

Professional Services (3 agents)

Management Layer (4 agents)

Helper Script

scripts/dl-api.py — Python script (stdlib only) for all API operations.

Commands:

Examples:

# Check health
python3 {baseDir}/scripts/dl-api.py health

# List all agents
python3 {baseDir}/scripts/dl-api.py agents

# Generate sales outreach
python3 {baseDir}/scripts/dl-api.py run sales_outreach '{"company":"Tesla","role":"VP Engineering"}'

# Write a business plan
python3 {baseDir}/scripts/dl-api.py run business_plan '{"business_idea":"AI-powered pet care app","market":"pet owners 25-45"}'

# Generate SEO content
python3 {baseDir}/scripts/dl-api.py run seo_content '{"keyword":"remote work tools 2026","content_type":"pillar"}'

# Run a batch of tasks
python3 {baseDir}/scripts/dl-api.py batch {baseDir}/examples/batch.json

Batch Processing

Create a JSON file with multiple tasks:

[
  {"agent": "sales_outreach", "inputs": {"company": "Stripe", "role": "CTO"}},
  {"agent": "ad_copy", "inputs": {"product": "AI CRM", "platform": "facebook"}},
  {"agent": "seo_content", "inputs": {"keyword": "business automation", "content_type": "blog"}}
]

Run them all:

python3 {baseDir}/scripts/dl-api.py batch tasks.json

Multi-Agent Workflows

Chain agents together for complex operations:

Lead-to-Close Pipeline:

1. lead_gen → find prospects in target industry
2. market_research → understand their pain points
3. sales_outreach → personalized outreach sequence
4. proposal_writer → ready proposal when they respond

Content Engine:

1. market_research → trending topics in your niche
2. seo_content → long-form SEO blog post
3. content_repurpose → tweets, LinkedIn, newsletter from that post
4. social_media → platform-native posts with CTAs
5. ad_copy → paid promotion copy

Client Onboarding:

1. doc_extract → pull data from client documents
2. data_entry → structure into your system format
3. bookkeeping → set up financial tracking
4. crm_ops → create CRM records

API Details

• Base URL: https://bitrage-labour-api-production.up.railway.app
• Universal Endpoint: POST /v1/run
• Agent List: GET /agents
• Health Check: GET /health
• Docs: GET /docs
• LLM Providers: OpenAI (GPT-4o), Anthropic (Claude), Google (Gemini), xAI (Grok)
• QA: Every output passes through QA verification before returning

Environment Variables

Important Notes

• Each agent call takes 3-15 seconds (real LLM inference, not cached)
• All outputs include QA verification status
• The provider field is optional in all agent inputs — omit to use the server default
• Management agents (context_manager, qa_manager, etc.) are internal and run automatically during pipelines
• Batch mode processes tasks sequentially to respect rate limits
• All agent outputs are structured JSON with consistent schema