Mino · Series
Codename: Veil
Build a document anonymisation tool, one step at a time
A daily mini-series where Dielis uses AI to build a tool that anonymises legal documents from scratch. One-minute episodes, from an empty screen to a working tool.
38 episodes · 36 min · with Dielis · New episode every weekday
-
0:36 1 What we're building: an AI redaction tool -
0:39 2 Create a GitHub account -
0:52 3 Install VS Code -
0:52 4 Install Claude Code in VSCode -
0:53 5 Install Python -
1:14 6 Your first Python file -
1:20 7 Make a project folder -
1:20 8 Set up a Python virtual environment -
0:48 9 Install our first package -
1:03 10 Read a PDF in Python -
1:04 11 Install git -
1:00 12 Your first Git commit -
1:00 13 Create a GitHub repo -
1:18 14 Add a .gitignore file -
0:56 15 Push to GitHub -
0:58 16 Your first redaction -
0:46 17 Save the result to a file -
0:53 18 Module one recap -
0:55 19 What is regex (and asking AI to explain it) -
1:03 20 Vibe-code your first detection -
0:52 21 Understand the code Claude wrote -
0:40 22 Generate a harder test document -
0:58 23 Detect license plates and the first edge case -
0:47 24 More patterns in one prompt -
0:53 25 Refactor the code with Claude -
0:53 26 Structured detection output -
0:59 27 The limit of regex -
0:55 28 Choose a tool and install spaCy -
0:46 29 Detect names with spaCy -
0:51 30 Where spaCy fails -
0:45 31 Combine regex and spaCy -
0:49 32 Module two recap -
1:16 33 What is an LLM -
0:45 34 Why local, why now -
0:36 35 Install Ollama -
1:52 36 Choose a model -
0:46 37 Download and run the model -
1:26 38 Connect Python to Ollama