llmtrim 0.1.6

Static, deterministic LLM prompt/payload compressor — cut input tokens 30-90% with zero extra model calls
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

//! Compression stages.

pub mod cache;
pub mod dedup;
pub mod hygiene;
pub mod image;
pub mod jsoncrush;
pub mod ngram;
pub mod ngram_sa;
pub mod output;
pub mod retrieve;
pub mod serialize;
pub mod sizing;
pub mod skeleton;
pub mod tool_schema;
pub mod toolout;
pub mod tools;

pub use cache::CacheStage;
pub use dedup::DedupStage;
pub use hygiene::HygieneStage;
pub use image::ImageStage;
pub use jsoncrush::JsonCrushStage;
pub use ngram::NgramStage;
pub use output::OutputControlStage;
pub use retrieve::RetrieveStage;
pub use serialize::SerializeStage;
pub use skeleton::{MinifyCodeStage, SkeletonStage};
pub use toolout::ToolOutputStage;
pub use tools::ToolStage;