AgentSidekick

Context Compression

Shrinking prompts and history to fit more into fewer tokens. As of 2026-11-21, 54 indexed products offer it (28 with full profiles), in 6 forms: 49 MCP servers, 1 Agent project, 1 Agent skill, 1 Library. 17 were published in the last 90 days, against 17 in the 90 days before. MemOS has the most GitHub stars, 52% of the total across 51 products with their own repository.

CapabilitySegment label: Lagging the catalogue DER

Available as

All in the directory

Profiled products grouped by product form. Within each form: own-repository products by GitHub stars, then products in a subfolder of a larger repository (they show that repository's stars), up to 10 each.

MCP servers 49 indexed · 23 profiled

Protocol-level reach: 36 of the 45 active platforms AgentSidekick tracks document support for MCP. DER Which platforms

Context Compression: MCP servers by GitHub stars
#ProductPublisherCapabilities GitHub stars Last push Save and compare
1sqzPre-injection context compression for coding agents. Zero LLM calls, zero telemetry, offline-safe.ojuschugh16819d ago
2EntrolyContext Assurance for AI agents: context engineering, AI cost optimization, verification.juyterman1000488today
3Code Context EngineIndex your codebase. AI searches instead of re-reading files. 94% token savings.ai-elara4672mo ago
4mcptoonMCP tools + agent skills in one zero-dependency CLI: 71,929 -> 581 tokens (-99.2%, measured).activeing1234413d ago
5MCPProxyLocal-first MCP proxy with BM25 tool discovery, security scanning, quarantine & ~99% token savingssmart-mcp-proxy428today
6Context-EngineContext-Engine MCP - Agentic Context Compression SuiteContext-Engine-AI402today
7llmtrimMCP server and proxy that compresses LLM prompts, tool output, and replies to cut token cost.fkiene2769d ago
8TOMC MemoryCPU task-specific context preparation and explicitly saved local project notes through MCP.ZipengWu3652641d ago
99routerx 9RouterX (9R Next Go Generation): Single-binary Go AI router, LLM gateway & 20-40% token saver. Route Claude Code, Cursor & Codex across 155 providers, 28 OAuth & 248 free models (24 keyless) — OpenAI, Claude, Gemini, DeepSeek, Groq, Ollama & MCP.yanstd888258today
10trace-mcpCode graph MCP: 81 languages, 91 frameworks, 72.7% fewer PR-review tokens, comprehension at paritynikolai-vysotskyi204today

All 23 profiled MCP servers with this capability

Agent projects 1 indexed · 1 profiled

No compatibility evidence is collected for this form.

Context Compression: Agent projects by GitHub stars
#ProductPublisherCapabilities GitHub stars Last push Save and compare
1MemOSSelf-evolving memory OS for LLM & AI Agents: ultra-persistent memory, hybrid-retrieval, and cross-task skill reuse, with 35.24% token savings and DeepSeek Harness support.MemTensor12k1d ago

Agent skills 1 indexed · 1 profiled

Protocol-level reach: 26 of the 45 active platforms AgentSidekick tracks document support for the Agent Skills format. DER Which platforms

Context Compression: Agent skills by GitHub stars
#ProductPublisherCapabilities GitHub stars Last push Save and compare
1add-rtkInstall rtk token-compression proxy into agent containers. Routes Bash tool calls through rtk for 60–90% token savings on dev commands (git, cargo, pytest, docker, kubectl, etc.).nanocoai—today

Libraries 1 indexed · 1 profiled

No compatibility evidence is collected for this form.

Context Compression: Libraries by GitHub stars
#ProductPublisherCapabilities GitHub stars Last push Save and compare
1LLMLinguaPrompt compression methods that drop low-information tokens using a small model to shrink long contexts.microsoft6.8k12d ago

Plugins 1 indexed · 1 profiled

1 of these are Claude plugins, which Claude Code documents (protocol-level). DER

Context Compression: Plugins by GitHub stars
#ProductPublisherCapabilities GitHub stars Last push Save and compare
1magic-compactLossless context compression plugin for Claude Code & OpenCode.aerovato1802mo ago

Rules packs 1 indexed · 1 profiled

No compatibility evidence is collected for this form.

Context Compression: Rules packs by GitHub stars
#ProductPublisherCapabilities GitHub stars Last push Save and compare
1cursor-handbookOpen-source rules engine for Cursor IDE — 110 rules, agents, skills, commands that give AI permanent memory of your standards. Demo: 30%+ token savings, blocks hardcoded secrets, type-check instead of 100K-token test runs, consistent code across team.girijashankarj332mo ago

Which platforms these forms can reach

All platforms

Protocol-level: the platform documents support for the standard. No specific pairing has been tested.

Platforms publishers name

PlatformProfiled products naming it
Claude Code2
Cursor2
OpenCode1
Codex1

Integrations most often paired

Third-party services named by products with this capability, whole catalogue.

IntegrationProducts Share
GitHub23.7%
Kubernetes11.9%

Keywords that trigger this capability

Matched as whole words, ignoring case, in a product's name, description and repository topics.

KeywordProducts matched
context compression19
token reduction13
compress context0
prompt compression5
reduce tokens0
token savings17

Segment statistics

Whole catalogue (54 indexed products). Calculated 2026-11-21.

54DERIndexed products28 with full profiles
17DERLaunched, last 90 days17 in the 90 days before
9DERMomentum vs catalogue50 = same pace as the catalogue
52%DERLargest project's share of starsAcross 51 products with their own repository
0.352DERConcentration (HHI)About 3 equal-sized projects’ worth; highly concentrated
20SRCMedian starsPer product with its own repository
83%DEROpen sourceMarked open source; unknown counts as not

Lagging the catalogue. Mid-sized among capabilities, with launches growing more slowly than the catalogue’s. Size is ranked against all capabilities; momentum compares launches in the last 90 days with the 90 before, relative to the catalogue’s own ratio (62 or more = faster, 38 or less = slower; under 5 dated launches it is not measured). The HHI of 0.352 means stars are spread as if across about 3 equally popular projects — highly concentrated by the common 0.15 / 0.25 thresholds. Stars measure attention, not usage or quality.

Limits

  • Classification is by the keywords above. A product that offers this capability without using one of them is not counted; one that uses a keyword in another sense is counted wrongly. Methodology
  • Platform reach is protocol-level: it says a platform documents the standard, not that any product here was tested on it. Platforms named by publishers are their own claims.
  • Stars belong to repositories. Products in a subfolder of a large repository show that whole repository’s stars and are excluded from the concentration figures.
  • Launch counts use each source’s publication date; for auto-discovered repositories that is the repository’s creation date.

Page data last updated .