token-efficiency

[ACL 2026 Oral] "LightReasoner: Can Small Language Models Teach Large Language Models Reasoning?"

One zero-dependency MCP gateway CLI for every server and agent skill. Token optimization, tool discovery and context compression: 71,929 -> 581 tokens (-99.2%, measured), schemas stay out of context. One config for Claude Code, Codex, Cursor, every agent. 561KB, pure Python. | 零依赖 CLI：管所有 MCP 工具与技能，工具发现省 99.2% token。

The context layer for AI coding agents: less noise, fewer tokens, nothing lost. Formerly OMNI.

Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success but 65% Fewer Tokens

An open-source, local-first AI teammate and agent operating environment for turning intent into completed work—with tools, memory, permissions, and human oversight.

Dev tools, optimized for agents. Structured, token-efficient MCP servers for git, test runners, npm, Docker, and more.

An agentic memory database that cuts session tokens by 82–99%. One portable SQLite file — your agent's memory, anywhere.

🔥 Token-efficient JSON alternative for LLMs & agentic AI — same data, fewer tokens. Python · JS/TS · Rust · Go · C++

Token-efficient, local-first CLI tools for coding agents - compact Maven, npm/Node, and Go test output plus reusable development helpers.

HEWN 2.0 2026: AI Output Router for Precision Summaries & Polished Code

Token Cost Parity: Multilingual LLM Efficiency Analysis 2026

dsh-sieve: context engineering & token optimization plugin for DeepSeek Harness (DSH) — tool output filtering, context pruning, progressive skill disclosure. 36% smaller payload in offline replay. DSH 上下文管理与 token 优化节省插件。

A curated list of strategies, tools, papers, and resources for reducing LLM token costs and improving efficiency in production.

Your AI forgets every session. Give it a memory it can trust: local-first, review-gated agent memory — one SQLite vault for DSH, Claude Code, Codex, Cursor and any MCP client, with an immutable audit log and a token ledger. Open source, self-hosted, no cloud.

Agents can work. They cannot declare success. A governed coordination protocol for multi-agent AI work: eight contracts (task, resource, ownership, handoff, verification, evidence, approval, recovery) enforced in code, plus a reproducible conformance benchmark. Apache-2.0 code, CC BY 4.0 specification, zero dependencies, Node 18+.

Agent Dashboard: Visualization and analytics for Sessions and Quota Usage. Track, analyze, and optimize token usage across providers with heatmaps, cost tracking, token counting and quota resets..

降低Deepseek v4.1flash的Token消耗/Raising Token efficiency without lossing performance(Works for v4.1flash)

Token-efficient data serialization for LLM/AI. 50% fewer tokens than JSON, 93% better value/token. Rust, schema validation, LSP.

Verified code context for agents

The AI-native wire format for structured data. 100% comprehension on every frontier model. 50-92% fewer tokens than JSON. 43B+ lossless round-trips across 17 formats. Spec v3.5.1 Stable.