Cost Reduction Python Packages

llmtrim

Local proxy that compresses your LLM API requests so you pay less, with no change to the answers. Trims wasted tokens from prompts, history, tool output, and code before they're sent: -31% input / -74% output, measured live. Any provider, no extra model calls. Also an MCP server and embeddable library (Rust, Python, Ruby, Kotlin, Swift, JS/TS).

rtk-hermes

RTK plugin for Hermes — rewrites shell commands for 60-90% LLM token savings

kalibr

Stop overpaying to run your agents. Kalibr routes every request to lower-cost model and tool paths without degrading performance.

latent-gate

VL-JEPA inspired pipeline — compress images/text locally via Ollama, send compact payloads to any LLM API. Cut token costs by ~80%.

graphsift

graphsift: #1 Claude token saver & LLM token optimizer. AST dependency graph, BM25+graph ranked relevance, multi-tier context selection, 14 languages, tree-sitter parsing, 19-CLI output compression (86% avg). 80-150x token reduction, F1 0.85. Hybrid search, dedup, diff-aware trimming, cycle & dead code detection. MCP server. Agent memory, typed graph retrieval, context compaction, A2A protocol, temporal code graph.

thrifty-ml

Proxy-model library: replace per-row LLM calls with lightweight ML classifiers

agentsave

Save 30% on AI agent costs. One line of code. No accuracy loss.

python-token-killer

Minimize LLM tokens from Python objects — dicts, code, logs, diffs, and more.

thinkrouter

Cut LLM reasoning-token costs by 60% with one line of code.

llm-response-cache

SQLite-backed LLM response cache. Exact match + fuzzy match. Decorator API. Zero mandatory server dependencies.

llm-cost-optimizer

Automatically reduce LLM API costs. Routes to cheapest model that succeeds. Use kalibr instead.

llm-cost-reduction

Stop overpaying to run your agents. Kalibr routes every request to lower-cost model and tool paths without degrading performance.

fortress-optimizer

Cut your LLM API costs by 10-20%. Drop-in prompt optimization for Python.

langchain-kalibr

LangChain integration for Kalibr. Route every LangChain LLM call to lower-cost model paths without degrading performance.

tokenwise-optimizer

Smart Context Optimization for LLMs - Reduce tokens by 70-90%