Cover of Modal Made Crystal Clear
← Back to the series

Modal

Made Crystal Clear

By Christophe Paka · Updated August 28, 2026

A practical, visual guide to running Python, GPUs and AI workloads in the cloud, paying only while your code runs

Prefer it on your Kindle, or on paper? Same book, on Amazon.

Give feedback

Who this book is for

In plain words: Modal runs your Python code and AI models on cloud GPUs, and you pay only for the seconds your code is actually running.

  • Python developers and ML engineers who need GPUs now and then
  • Technical founders serving an AI model without a platform team
  • Not for: Not for you if you don't write Python. Try Hugging Face Made Crystal Clear.

Every GPU tutorial ends the same way: a machine you forgot to turn off. Modal sells the other unit, functions and seconds. This guide takes you from pip install modal to a deployed vLLM endpoint your OpenAI client code already speaks, through environments written as Python, cron jobs and web APIs that cost pennies a month, GPUs picked by arithmetic instead of spreadsheet, and Sandboxes for code an LLM wrote. It treats the catches as content: cold starts measured before the levers, and the break-even worksheet that says when per-second billing stops winning. Every price is dated and every vendor claim is attributed.

Pages
198
Level
Beginner to advanced
Updated
August 28, 2026

12 real pages. Click a page to enlarge it.

Page 9: where Modal sits
Page 19: the big schema
Page 21: modal's stack in layers
Page 39: the Image layer cake
Page 64: the GPU menu with approximate price per hour, derived from Modal's per-second rates
Page 67: the hardware safety net
Page 74: the commit/reload contract between two containers
Page 98: the four segments of a cold start and the lever that attacks each
Page 123: the same platform, two entry points
Page 124: the Sandbox lifecycle
Page 137: the break-even curve
Page 164: the reference card

Inside the book

  1. 01The End of Paying for Idle
  2. 02The Big Picture: How Modal Thinks
  3. 03Zero to Cloud in Ten Minutes
  4. 04Images: Your Environment as Python Code
  5. 05Python in the Cloud 24/7: Cron Jobs and Web Endpoints
  6. 06GPUs on Demand: Picking Hardware Without a Spreadsheet
  7. 07Volumes: Where Model Weights Live
  8. 08Ship an LLM Endpoint: vLLM in Production Shape
  9. 09Cold Starts, Solved Honestly
  10. 10Scaling Out: Map, Spawn, and Training Runs
  11. 11Sandboxes: A Safe Room for Untrusted Code
  12. 12The Money Chapter: What Modal Really Costs
  13. 13Production: Secrets, Regions, Budgets, and Sleep
  14. 14Your 90-Day Path and the One-Page Cheat Sheet
  15. 15Resources: The Short List Worth Your Time

You'll leave with

  • Understand what happens between your decorator and a billed second, and why idle costs zero
  • Write your environment as Python, with model weights baked in at build time
  • Ship a cron job and a locked web endpoint for pennies a month, no GPU required
Christophe Paka, author of Modal Made Crystal Clear

About the author

Christophe Paka writes every book on the shelf. He spent years in technical support at Salesforce and Microsoft, where the job was to make complex systems clear to customers who needed them working the same day. He has since built and run software as a founder and lead developer, worked as a Salesforce consultant, and grown a YouTube channel on tech careers past one million views.

This book was last updated on August 28, 2026.

Connect on LinkedIn →

For teams

Put Modal on every desk at your company.

One licence, every book in the series, one link for everyone. From $250 a year.

See the team licences

Sources

Every source cited in Modal Made Crystal Clear. Each number matches the [n] marker next to that link in the book's "Go deeper" lists, so you can trace any claim back to its origin.

  1. [1]Modal homepage
  2. [2]Modal customer stories
  3. [3]Latent Space: "Truly Serverless Infra for AI Engineers"
  4. [4]Erik Bernhardsson: "What I have been working on: Modal"
  5. [5]Coiled: a GPU-accelerated serverless alternative to AWS Lambda
  6. [6]Hacker News: practitioners on Modal
  7. [7]Modal pricing
  8. [8]Functions guide
Show all 53 sources
  1. [9]Images guide
  2. [10]Autoscaling guide
  3. [11]Hello world example
  4. [12]Web endpoints guide
  5. [13]Modal changelog
  6. [14]Edlitera's Modal tutorial
  7. [15]modal.com/docs/guide/images
  8. [16]modal.com/docs/guide/existing-images
  9. [17]modal.com/docs/examples/hello_world
  10. [18]Scheduling remote cron jobs
  11. [19]The Server primitive
  12. [20]GPU acceleration guide
  13. [21]Billing guide
  14. [22]GPU cold starts and VRAM math
  15. [23]Volumes guide
  16. [24]Cloud bucket mounts
  17. [25]Dicts
  18. [26]Queues
  19. [27]Deploying a vLLM server on Modal
  20. [28]Modal docs: Functions
  21. [29]Modal docs: Web endpoints
  22. [30]Modal docs: Volumes
  23. [31]Modal docs: GPU acceleration
  24. [32]Modal: Cold start performance
  25. [33]Modal: Memory snapshots
  26. [34]Modal: GPU memory snapshots
  27. [35]Modal: GPU snapshot example
  28. [36]Modal: engineering talk on the container stack
  29. [37]Scaling out (Modal docs)
  30. [38]Substack case study (Modal)
  31. [39]Sandboxes guide
  32. [40]Sandboxes product page
  33. [41]The safe_code_execution example
  34. [42]JavaScript and Go SDKs
  35. [43]Security guide
  36. [44]Budgets
  37. [45]Introl: serverless GPU platforms compared
  38. [46]GPU Tracker: serverless GPU comparison
  39. [47]Secrets guide
  40. [48]Region selection
  41. [49]Developing and debugging
  42. [50]OpenTelemetry integration
  43. [51]Managing deployments
  44. [52]Web Function URLs
  45. [53]modal.com

New books, in your inbox

One short note when a new Made Crystal Clear guide ships.
No spam, unsubscribe anytime.