Skip to content
Optimal Lab

Tools

See how the tools around the model actually move: agents, MCP, and workflow builders.

Sessions

Go in order. Each session is one idea you can drag, step and break.

01Coding agents

1 live · 4 coming

Tools that read and change your code.

  1. Interactive

    Live

    A coding agent's loop

    Watch search, edit, and a check repeat until the test passes.

    Open session →

  2. Interactive

    Coming soon

    Terminal coding agents

    Agents like Claude Code that run commands, edit files and run tests.

  3. Interactive

    Coming soon

    Rules and context files

    AGENTS.md, rules and memory files that steer an agent.

  4. Interactive

    Coming soon

    Reviewing agent changes

    Diffs, tests and sandboxes before you trust the output.

  5. Interactive

    Coming soon

    Background and cloud agents

    Hand off a task and get a pull request back.

02Protocols and tool servers

2 live · 2 coming

Standard ways to plug tools into models.

Already live in other tracks: Call a tool, then answer

  1. Interactive

    Live

    MCP: tools the model can call

    A server lists tools. The app forwards the one the model picks.

    Open session →

  2. Interactive

    Live

    Building an MCP server

    Expose your own data and actions to any MCP client.

    Open session →

  3. Interactive

    Coming soon

    Turning an API into tools

    Generate tool definitions from an OpenAPI spec.

  4. Interactive

    Coming soon

    Agent-to-agent protocols

    How agents from different vendors hand work to each other.

03Workflow builders

1 live · 3 coming

Visual and code-first automation.

Already live in other tracks: One item through a workflow

  1. Interactive

    Live

    An n8n workflow, node by node

    Cross a threshold and see the run switch from email to Slack.

    Open session →

  2. Interactive

    Coming soon

    Zapier and Make

    No-code automation, and where it runs out.

  3. Interactive

    Coming soon

    LangGraph

    Agents as graphs of steps with saved state.

  4. Interactive

    Coming soon

    Temporal

    Code-first durable workflows that survive crashes.

04Model APIs and SDKs

1 live · 4 coming

Calling models from your code.

  1. Interactive

    Coming soon

    Model provider APIs

    OpenAI, Anthropic and Gemini: messages, tools and streaming compared.

  2. Interactive

    Coming soon

    AI SDKs

    Vercel AI SDK and friends: one interface over many providers.

  3. Interactive

    Coming soon

    LangChain and LlamaIndex

    What frameworks give you, and what they hide.

  4. Interactive

    Coming soon

    Ollama and local runtimes

    Run open models on your laptop behind a local API.

  5. Interactive

    Live

    Model routers

    OpenRouter-style services: one key, many models, automatic fallback.

    Open session →

05Data and retrieval tools

0 live · 3 coming

Where embeddings and documents live.

Already live in other tracks: Embeddings and similarity search

  1. Interactive

    Coming soon

    Vector databases

    pgvector, Pinecone, Qdrant and Weaviate compared.

  2. Interactive

    Coming soon

    Search engines

    Elasticsearch and OpenSearch for keyword and hybrid search.

  3. Interactive

    Coming soon

    Document parsers

    Getting clean text and tables out of PDFs and slides.

06LLMOps tools

0 live · 4 coming

Tracing, evals and prompts in production.

Already live in other tracks: Evals: is the new prompt better?

  1. Interactive

    Coming soon

    Tracing tools

    Langfuse, LangSmith and similar: see every call and its cost.

  2. Interactive

    Coming soon

    Eval tools

    promptfoo, Braintrust and friends: run test sets on every change.

  3. Interactive

    Coming soon

    Guardrail tools

    Libraries that check inputs and outputs against rules.

  4. Interactive

    Coming soon

    Prompt management

    Store, version and roll out prompts outside your code.

07CI tools

1 live · 1 coming

Automation around your repository.

Already live in other tracks: A CI/CD pipeline, stage by stage

  1. Interactive

    Live

    GitHub Actions on every push

    Follow a workflow file from the push event to the job log.

    Open session →

  2. Interactive

    Coming soon

    AI checks in CI

    Run AI review and eval suites on every pull request.

In production

Tools you can use

21 technologies across 6 live sessions. Same names in any language — pick a session to see how the idea maps to the real world.

Services, platforms, protocols and standards — not one stack.

Roadmap

Coming soon

A step-by-step path through the track, with sessions and practice at each stage.

  1. Stage 1

    Coding agents

  2. Stage 2

    MCP and protocols

  3. Stage 3

    Workflow builders

  4. Stage 4

    Model APIs and data tools

  5. Stage 5

    LLMOps

Projects

Coming soon

Build something real to lock in the skill.

  • Connect one MCP tool
  • An n8n flow with a branch
  • A GitHub Action that runs tests