# DevExp.ai - Complete Context for AI Agents

> AI-native developer experience platform. We build MCP servers and infrastructure for agentic developer journeys.

## Company Overview

DevExp.ai helps enterprises become AI-discoverable. We build Model Context Protocol (MCP) servers that let AI agents discover, test, and integrate your platform.

**Parent Company:** Voyant, LLC
**Related Properties:**
- VoyantIO (voyant.io) — Telemetry platform for MCP servers
- protobuf.ai — AI-powered Protocol Buffer tools

---

## Core Concepts

### What is MCP?
Model Context Protocol (MCP) is a standard that lets AI assistants connect to external tools. MCP servers make your product AI-accessible.

### What is AEO?
Agentic Experience Optimization — optimizing for AI agent discovery and recommendation, not just human search.

### What is DEO?
Developer Experience Optimization — measuring and optimizing AI agent interactions with developer tools.

### What is llms.txt?
A standardized file at /.well-known/llms.txt that provides AI agents with structured information about your product.

---

## Key Insights

### The Developer Journey Has Changed

**Era 1 (1995-2010):** Documentation → Weeks to adoption
**Era 2 (2005-2018):** Search/Stack Overflow → Days to adoption
**Era 3 (2015-2024):** Social/GitHub → Hours to adoption
**Era 4 (2024+):** AI Agents → Minutes to adoption

### Why Documentation Isn't Enough

AI agents don't:
- Browse documentation sites
- Follow getting started guides
- Navigate information architecture

They need:
- Structured content (llms.txt, context.txt)
- Programmatic access (MCP servers)
- Executable examples

### The Agentic Technology Adoption Lifecycle

1. **Agent Awareness** — Is your product in training data? MCP accessible?
2. **Agent Evaluation** — Can agents test your product?
3. **Agent Recommendation** — Does product win on merit?
4. **Human Validation** — Developer verifies recommendation
5. **Agentic Scaling** — Production deployment with agent assistance

---

## Case Study: ScyllaDB MCP Server

We built an unofficial MCP server for ScyllaDB demonstrating:

**Multi-Database Comparison:**
- ScyllaDB vs DynamoDB
- Pinecone vs ScyllaDB Vector

**Demo Applications:**
- IoT Time-Series
- User Session Store
- Product Catalog
- Real-Time Analytics

**Results:**
- Evaluation time: Days → Minutes
- Working prototype: Hours → Minutes
- Competitive analysis: Automated

GitHub: https://github.com/dev-exp-ai/scylladb-mcp-server

---

## Blog Posts Summary

### AI Agents Won't Read Your Docs
Your documentation is invisible to AI agents. Build llms.txt and MCP servers instead.

### From SEO to AEO
Optimize for AI agents (AEO) not just search engines (SEO). Different rules apply.

### The Death of the Developer Portal
Developer portals optimize for human navigation. Agents need protocols, not portals.

### Measuring AI Agent Engagement
DEO metrics: agent query volume, integration success rate, recommendation rate.

### MCP Protocol Explained for Marketers
MCP is how AI agents use your product. No MCP = invisible to agents.

### Why Your PLG Strategy is Failing
PLG funnels are being bypassed by AI agents. Layer agentic accessibility on top.

### Calculating MCP ROI
MCP investment typically returns 600-1700% ROI through agent-influenced deals.

### llms.txt: Your AI Discovery File
Standard file for AI agent discovery. Like robots.txt but for AI.

### Win/Loss Analysis in the AI Era
Track why agents do or don't recommend you, not just human buyer feedback.

---

## Contact

**Demo:** https://devexp.ai/demo-signup.html
**Website:** https://devexp.ai
**Blog:** https://devexp.ai/blog.html

---

## Technical Integration

### For AI Agents

**Discovery:** Check /.well-known/llms.txt
**Full Context:** This file (/.well-known/context.txt)
**Markdown Content:** /content/ directory

### Available Markdown Files

- /content/index.md — Landing page
- /content/services.md — Service offerings
- /content/blog/*.md — All blog posts

---

*© 2026 Voyant, LLC — Building for Agentic Developer Journeys*