optiflow-mcp
A Claude Code plugin that merges token optimization and compression tools into a single MCP server, enabling context-aware token management, session analytics, and compact-continue checkpoints for efficient AI interactions.
README
optiflow-mcp
A single Claude Code plugin, one MCP server, one repository — not a wrapper around separately-installed tools. It genuinely merges the real source of two upstream projects into its own codebase:
- token-optimizer-mcp's
76 real
smart_*/analytics/dashboard-monitoring MCP tools, plus itsPreToolUse/PreCompactenforcement hooks, ported directly intosrc/optimizer/— nonpxinstall, no separate process. - headroom's actual Rust
compression core (
SmartCrusher) forked intonative/headroom-core/, feature-stripped, and compiled to real WebAssembly (native/headroom-wasm/) loaded directly by the Node plugin — the genuine Rust algorithm, not a rewrite. headroom's CodeCompressor and Kompress ML model are reimplemented natively in TypeScript instead (src/native/). - Five modules of its own that neither upstream tool covered: chop-style
Bash/CLI-output interception, session-report transcript analytics, a
statusline context meter,
/optiflow:compact-continuesession-handoff checkpoints, and TOON conversion for large JSON/CSV/YAML payloads.
Both upstream projects are still referenced as git submodules under
vendor/ for provenance/license text, but nothing under src/ or native/
imports from vendor/ — see
docs/ADR/0002-real-merge-not-orchestration.md
for why this replaced the original orchestration-wrapper design (v1), and
docs/architecture.md for the full authority map and
locked decisions.
Status
Both major merges are done and wired: the 76-tool + enforcement-hook merge from token-optimizer-mcp, and headroom's compression core (WASM SmartCrusher
- TS CodeCompressor/Kompress) actually called from the shipped pipeline, not
just built in isolation.
npm run build,npm test, andnpx tsc --noEmitall pass — seedocs/modules.mdfor per-module detail and current test count. Building from source needs a Rust toolchain (rustup, thewasm32-unknown-unknowntarget,wasm-pack) in addition to Node; a plugin install does not (the compiled.wasmis committed, same reasoning asplugin/dist/).
Getting started
Published at github.com/kristijankopacevic/optiflow-mcp.
Verified end-to-end via a local-path marketplace (no gh auth needed for
that):
- In a Claude Code session, add this repo as a marketplace and install the
plugin:
/plugin marketplace add C:\path\to\optiflow-mcp /plugin install optiflow@optiflow - Check your environment:
Reports Node/npm versions, config resolution, andoptiflow doctorghpresence/auth. - Activate the statusline (optional, opt-in — installing the plugin alone
does not touch your settings.json):
Backs up youroptiflow install --statusline~/.claude/settings.jsonbefore writing, writes atomically, and refuses to overwrite a different existing statusline without--force.optiflow uninstallreverses it. Seedocs/statusline-manual-setup.mdfor details (including doing it by hand) anddocs/modules.mdfor the rest of the modules.
See docs/ for architecture, configuration, and per-module
reference depth beyond this quick start.
License
MIT for optiflow-mcp's own code (see LICENSE). This repository
also copies in source from MIT-licensed (token-optimizer-mcp) and
Apache-2.0-licensed (headroom) upstream projects — see
THIRD_PARTY_LICENSES.md and
NOTICE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
E2B
Using MCP to run code via e2b.