Running Inception as a service, Foundation Model Platforms

Mercury Agency Implementation, High-Speed AI Agents and Automation

Learn how to build and monetize AI agent services using Inception's Mercury models, including voice automation, browser-use workflows, and real-time customer support systems. This course covers API integration via LangChain and LiteLLM, cost optimization through token caching and Mercury Router, and productized delivery models for agencies scaling AI projects.

Open the decision record for Inception

What does running Inception for clients commit you to?

Published figures for this service. Blank fields are not published.

Monthly tool cost
Vendor cost basis is usage-based: Mercury 3 at $0.10 per 1M input and $0.35 per 1M output during the launch 50% discount. Exact monthly spend depends on the agency's own token volume, which is not published and must be supplied by the agency.
Time to first value
Hours to first value (published time_to_value); setup_complexity is medium
Payback
Not modeled
Guided implementation
8 hours

Is Inception worth running as a client service?

The evidence supports Inception as a fast, OpenAI-compatible model API fit for latency-sensitive agency builds, but it is a raw API with no agency dashboard or white-label layer. Whether the engagement is profitable depends on client price, labor, and usage volume, which are not supplied.

An agency-fit judgement for reselling this service. It is separate from the tool description on the decision record.

Before you start

What has to be in place before the first client engagement.

Tools and subscriptions

  • Inception Platform API key for the chat completions endpoint
  • OpenAI-compatible client library or SDK (AISuite, LiteLLM, or LangChain)
  • Optional distribution route via OpenRouter, Baseten, or Vercel
  • Structured-output schema definitions for Mercury Decide classification jobs
  • Mercury Router configuration for quality/speed/cost routing

People and inputs

  • Developer capacity to call and integrate an API (setup_complexity: medium)
  • A latency-sensitive client workflow to benchmark (voice, browser-use, or support)
  • Test prompt sets and acceptance criteria for model evaluation
  • Usage-monitoring setup to reconcile token spend against vendor rates

Included with the course

7 working documents for delivering this service.

  • Mercury Model Selection Worksheet for Client Projectsworksheet
  • Voice Agent Deployment SOP with Sub-170ms Latency Targetssop
  • Token Caching Strategy Checklist for Cost Reductionchecklist
  • Mercury Router Configuration Template for Multi-Workload Optimizationtemplate
  • LangChain Integration Guide for Agency Developersguide
  • Customer Support Automation Pricing Calculatorworksheet
  • Mercury Decide Classifier Setup for High-Volume Triagesop

Listed by name. These documents are not yet published as individual downloads.