AI Knowledge Filler prompt/widget to structure LLM output

Ask Claude for a note — get a wall of prose. Then spend 5 minutes adding YAML frontmatter, fixing headings, linking related notes. Every single time.

There’s a better way. The whole flow runs from a phone widget:

The result lands directly in your vault — complete .md file with validated metadata, heading hierarchy, WikiLinks, domain classification. No post-processing.


What gets enforced automatically

Without AKF With AKF
Random or missing frontmatter 8 required fields, validated
Inconsistent tags 30+ domain taxonomy
No related notes WikiLinks generated
Different structure every time Same format, every LLM

Two ways to use it

No setup — Upload system_prompt.md to a Claude Project. Start generating immediately.

CLI / widget (free Groq tier):

pip install ai-knowledge-filler
export GROQ_API_KEY="..."
akf generate "Api rate limiting guide"

GitHub: GitHub - petrnzrnk-creator/ai-knowledge-filler: Transform any LLM into a deterministic knowledge base generator. Universal system for structured Markdown with YAML metadata.


Not useful if your vault is freeform notes — this adds structure overhead you don’t need.

How are you currently handling AI-generated notes? Reformatting manually, or found a better workflow?


Changed title to be more specific and include the name of the thing being shown (was “How to handle AI-generated content in your vault?”).

1 Like