Retrieve
Your existing RAG stack finds candidate messages, chunks, notes, or documents.
Trajbl
Keep the evidence. Drop the noise.
Trajbl sits between retrieval and the LLM. It receives a retrieved context pool, keeps complete evidence sentences, and builds a compact packet that the model can answer from.
+72%relative F1 lift vs LLMLingua2controlled long-context benchmark, similar token budget
~88%context token reductionmeasured against full-context RAG reference
68%RAG quality retainedwhile sending a much smaller context packet
0GPU/model dependencyTrajbl compression is CPU-first and model-free
Pipeline role
Trajbl is not trying to replace retrieval.
Good compression needs a good input pool. The fair production setup is: existing retriever top50 to Trajbl to LLM. That is why the benchmark compares Trajbl and LLMLingua2 over the same vector, hybrid, and BGE-reranked retrieval setups.
Rank or fuse
Optional vector, hybrid RRF, or BGE reranking creates a stronger top50 context pool.
Compress with Trajbl
Trajbl keeps a small source-traceable packet of complete evidence sentences.
Answer
The LLM reads less context while preserving the facts most likely to answer the question.
Controlled benchmark signal
Trajbl beat LLMLingua2 on quality-per-token.
Full-context RAG remains the quality ceiling. Trajbl's job is to preserve much of that answer quality while sharply reducing the context packet. In the measured long-context comparison, Trajbl retained more quality than LLMLingua2 at a similar token budget.
Baseline comparison
Model-free Trajbl against model-based compressors.
Trajbl remained ahead of LLMLingua2 and LongLLMLingua-style compression in the measured long-context signal.
Evidence integrity
Selection at sentence level keeps the packet readable.
Whole sentences
Selected evidence remains readable and verbatim instead of becoming disconnected token fragments.
Context-aware
The selector balances relevance, answer cues, and evidence coverage while keeping implementation details proprietary.
Source-linked
The retained packet can preserve a trace back to the documents or messages that supplied it.
CPU-first
Trajbl is a local preprocessing layer and does not require a GPU or hosted compression model.