Thirty days. From your first toolset.execute_action(...) to a script that paginates, retries, filters, and acts. Same six prompts you saw on day 1 — rate yourself again.
Where do I go next?
AI Beginner, if you haven't done it. The shape stays the same — write Python that calls a thing — but the thing is a language model instead of a Composio tool. You'll send prompts, parse JSON responses, and learn structured output. The kit you have here applies directly: helpers, retries, error handling, dispatch, batch.
If you've done both Beginner tracks, you have everything needed to write your own real-world automations: Python primitives, tool calls, idempotency, retry, and (after AI-Beg) LLM calls. The Intermediate and Advanced tracks of all three series unlock more depth — but the foundation is what matters most.
24 small Python scripts across four weeks plus the integration synthesis. Your kit:
| Capability | Lessons |
|---|---|
| Read tool responses | L1-L7 (week 1: tool calls, list/dict responses, filter, args, errors) |
| Write to the world | L8-L13 (send, create, append, update, idempotency, dedup) |
| Compose tools across services | L15-L21 (read-then-write, filter-then-act, time windows, error chains, logging, synthesis) |
| Production patterns | L22-L28 (retry, pagination, helpers, batch, dispatch, integration) |
Any automation script you'll read or write from here is a composition of these.
Deferred to Auto Intermediate (when v1 expands beyond 3 tracks):
Auto Advanced adds: durable workflows, idempotency keys at scale, cross-service consistency, replay, observability dashboards.
None of these are required to write your first dozen useful automations. You have enough.
→ AI Beginner — LLM calls, prompting, the four task verbs (summarise / infer / transform / expand), structured output, multi-turn, system prompts. Pairs naturally with what you just learned: classify the email body with an LLM before deciding which channel to dispatch to.
→ Or write your own first automation. Pick a pain — "send me a summary every morning", "log specific Slack messages to a Sheet", "alert me when X happens". The kit you have can make it.
Rate the prompts below as honestly as you did on day 1.
Thirty days. From your first toolset.execute_action(...) to a script that paginates, retries, filters, and acts. Same six prompts you saw on day 1 — rate yourself again.
Where do I go next?
AI Beginner, if you haven't done it. The shape stays the same — write Python that calls a thing — but the thing is a language model instead of a Composio tool. You'll send prompts, parse JSON responses, and learn structured output. The kit you have here applies directly: helpers, retries, error handling, dispatch, batch.
If you've done both Beginner tracks, you have everything needed to write your own real-world automations: Python primitives, tool calls, idempotency, retry, and (after AI-Beg) LLM calls. The Intermediate and Advanced tracks of all three series unlock more depth — but the foundation is what matters most.
24 small Python scripts across four weeks plus the integration synthesis. Your kit:
| Capability | Lessons |
|---|---|
| Read tool responses | L1-L7 (week 1: tool calls, list/dict responses, filter, args, errors) |
| Write to the world | L8-L13 (send, create, append, update, idempotency, dedup) |
| Compose tools across services | L15-L21 (read-then-write, filter-then-act, time windows, error chains, logging, synthesis) |
| Production patterns | L22-L28 (retry, pagination, helpers, batch, dispatch, integration) |
Any automation script you'll read or write from here is a composition of these.
Deferred to Auto Intermediate (when v1 expands beyond 3 tracks):
Auto Advanced adds: durable workflows, idempotency keys at scale, cross-service consistency, replay, observability dashboards.
None of these are required to write your first dozen useful automations. You have enough.
→ AI Beginner — LLM calls, prompting, the four task verbs (summarise / infer / transform / expand), structured output, multi-turn, system prompts. Pairs naturally with what you just learned: classify the email body with an LLM before deciding which channel to dispatch to.
→ Or write your own first automation. Pick a pain — "send me a summary every morning", "log specific Slack messages to a Sheet", "alert me when X happens". The kit you have can make it.
Rate the prompts below as honestly as you did on day 1.
Create a free account to get started. Paid plans unlock all tracks.