Engineering
Context engineering: what goes into your agent’s window
Instructions, input, tool definitions and tool results all compete for the same context window. Four habits that keep it full of what matters.
2 min read
Blog
Engineering
How Agent200 works under the hood: execution, tool calling and the infrastructure around AI agents.
Engineering
Inside run(): how the tool-calling loop works
Making a tool available is not the same as calling it. A walk through one run() request: the model, the instructions, the tools, and…
2 min read
Engineering
Every agent rebuilds the same plumbing. It doesn’t have to.
Integrations, OAuth, tool schemas, credentials and a bill per provider: the work around an AI agent repeats from project to project. Here is that…
2 min read
Engineering
Connectors and tools: what your model actually sees
A connector is access to a service. A tool is a capability the model can call. The difference shapes what your model sees, and…
2 min read
Engineering
Designing agents for many users from day one
A demo has one user. A product has thousands, each with their own accounts. How per-user connections work, and four rules for your side.
2 min read
Engineering
Tool calling, from first principles
A model cannot call anything. It writes a structured request and something else executes it. How the exchange works, and where Agent200 sits in…
2 min read
Engineering
Agents that answer with sources
A claim asks for trust; a cited fact earns it. Which tools return references, how to ask for citations, and how to design answers…
2 min read
Engineering
RAG Worked Because Your Data Sat Still
For years RAG was the default way to connect AI to data, and it worked because the data didn't move. Why that stopped being…
6 min read
Engineering
How to Connect AI Agents to Email
Connecting an AI agent to email is a six-layer problem, and most teams only budget for two. Gmail API, Microsoft Graph, IMAP, parsing services…
4 min read
Engineering
Email Context for AI Agents: Build or Buy?
RAG works on documentation because it has clean structure and no concept of time. Email has neither. The six layers of email context, three ways to get them, and what each costs.
3 min read
200 OK
Build the agent.
Access everything it needs.
Bring the agent you already have. Agent200 provides and executes its external capabilities.