People ask me "what does M87 do?" and I never give the same answer twice. Not because I don't know, but because it depends on who's asking.
A founder with an idea needs something different from a CTO with a team. A startup that lost its developer needs something different from a company building a SaaS from scratch. We've done all of it. Here's how each one actually works.
A. The Founder With a Vision
A non-technical founder comes to us with an idea. They have the market knowledge, the connections, maybe some funding. What they don't have is an engineering team.
We start with scoping. What's the smallest version of this product that proves the idea? We architect the system, design the UX, build it, launch it, and maintain it. The founder focuses on their market. We handle everything technical.
One of our clients was a VP of Technologies & Business Development who had deep relationships across law firms, insurance companies, and car dealerships. He wanted an AI platform that could turn his clients' documents into customer support and lead generation. We scoped it, built it, launched it. The platform now serves 10+ enterprise clients.
Another was an entrepreneur testing the AI entertainment niche. He had a concept for an AI-powered tarot reading game. We built a full interactive experience with multiple AI reader characters, real-time card drawing, and persistent memory across sessions. Ten weeks from concept to product.
This is our most common engagement. Founder has vision, M87 builds the product.
B. The Team That Lost Its Developer
A team has a running product, real users, and a codebase nobody understands. The original developer left, or the agency that built it moved on. The code works, mostly, but nobody can maintain it.
We take over. We read the codebase, understand the architecture, stabilize what's broken, and pick up where the last team left off. Bug fixes, feature requests, deployments. We become the engineering team.
A team of crypto veterans bought a launchpad codebase that arrived mothballed. No deployment, no documentation, no one who understood it. We brought it back to life, deployed it across multiple blockchain networks, extended it with new features, and maintained it for a year and a half. They focused on deal flow. We ran the engineering.
C. The CTO Who Wants AI-Native Engineering
A CTO at a funded company brings us in because they want their team to build differently. Not just "use Copilot" - actually restructure how engineering works with AI.
We embed in the team. We build AI-native development pipelines, set up agent workflows, and demonstrate what's possible by shipping real features using these tools. The team sees the output, learns the patterns, and adopts them.
Our engagement with a cloud infrastructure company started exactly like this. The CTO wanted to accelerate AI-native practices. We made a critical algorithm self-maintaining using AI, then stayed to build the full production pipeline. We're now embedded inside the team.
D. The Full SaaS Build
Sometimes the project isn't an MVP. It's a full product. Auth, billing, admin dashboards, real-time features, AI integrations, the whole thing. Multi-year, production-grade.
We build these with our proprietary AI development pipelines. The same agentic workflows we use internally, applied to client work. The result is senior-quality output at a pace that traditional teams can't match.
Two chess grandmasters came to us with a unique education methodology they wanted to turn into software. We built the entire platform over two years. Interactive minigames, AI coaching, achievement systems, audio narration, admin dashboards. A full gamified learning experience, ready for launch.
We also built Promptchainer as CTO for a founder. A visual AI workflow builder with a 13-package monorepo, custom graph execution engine, and async worker pool. It reached $2k MRR before the engagement ended.
E. The DevOps Tool
A client needs a specific developer tool built. Not a product for end users, but infrastructure tooling for their engineering team.
We figure out the needs together. What problem are we solving? What does the team's workflow look like? Then we design and build a tool that fits, and we maintain it.
A hardware startup needed a way to manage cloud infrastructure without giving developers direct access to the hosting provider. We built a YAML-driven CLI with a diff engine, upgrade pipeline, database operations, and a watch mode designed for LLM agents. The founder got the access control he wanted. The developers got a tool that enforces discipline.
F. AI-Native Staffing
A startup needs developers but can't justify senior rates. Or they need to move fast and don't have time to hire.
We put M87 dev agents in their Slack. AI-augmented developers that ship features, fix bugs, and maintain code. Every piece of output is hand-reviewed by a senior M87 engineer. The client gets senior-quality work at a fraction of the cost.
This is the most cost-effective way to work with us. The AI does the heavy lifting. The human guarantees the quality.
The Pattern
Every engagement is different, but the pattern is the same. You tell us what you need. We figure out the right way to deliver it. We ship. We stay as long as you need us.
No juniors. No handoffs. No BS.
