# XFMS — Model Source > Pick the right LLM for any task. Ranked shortlist with rationale across 8 evaluators. - **Type:** MCP server - **Install:** `agentstack add mcp-visionairyse-xfms` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [VisionAIrySE](https://agentstack.voostack.com/s/visionairyse) - **Installs:** 0 - **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml) - **Latest version:** 0.4.0 - **License:** MIT - **Upstream author:** [VisionAIrySE](https://github.com/VisionAIrySE) - **Source:** https://github.com/VisionAIrySE/XFMS - **Website:** https://xpansion.dev ## Install ```sh agentstack add mcp-visionairyse-xfms ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # XFMS — Xpansion Framework Model Source [](https://pypi.org/project/xfms/) [](https://pypi.org/project/xfms/) [](LICENSE) [](https://smithery.ai/servers/russ-mwq9/xfms) [](https://xpansion.dev) **Pick the right LLM for your task — without the Twitter vibes.** State what you're using the model for. XFMS aggregates evidence from eight independent benchmark sources, normalizes it onto a common scale, lets your intent decide which dimensions matter, and returns a ranked shortlist with plain-English rationale for every pick. XFMS is one module of the **[Xpansion Framework](https://xpansion.dev)** — a unified architecture for governing AI-assisted work. --- ## What this repository is A **thin Python client** and **command-line tool** for calling the hosted XFMS API at `xfms.xpansion.dev`. About 250 lines of code. It turns a one-liner into a ranked LLM shortlist. What this repository is **not**: the recommender engine, the score catalog, or the ingestion pipeline. Those run on the hosted service. The methodology behind every pick is published in full at [docs/methodology.md](docs/methodology.md) — every claim there maps to code that runs at request time, you just don't run it locally. --- ## What you say: > *"Fixing bugs in our Python codebase."* ## What you get: ``` Top picks: 1. 0.842 GPT-5.5 (openai/gpt-5.5) via OpenAI 2. 0.811 Claude Opus 4.7 (anthropic/claude-opus-4.7) via Anthropic 3. 0.798 Gemini 3.1 Pro Preview (google/gemini-3.1-pro-preview) via Google Inferred quality weights from your purpose: • structured_output_reliability 42.0% ← BigCodeBench, Aider Polyglot • instruction_following 28.0% ← LiveBench, Tau-Bench • factuality 20.0% ← MMLU, GPQA • coherence 10.0% ← LongBench ─── Explanation ─── Picked GPT-5.5: strong on structured output and instruction following — the two dimensions that dominate code-edit work. Beats Claude on Aider Polyglot and matches it on LiveBench reasoning, at roughly 60% of the per-token cost. ``` Want to see how the picks actually behave on your kind of query? Add `--ab`: ``` ─── A/B probe ─── Ran 5 test queries against the top picks. • GPT-4o-mini avg_latency=5579 ms total_cost=$0.00156 successes=5 • GPT-5.5 avg_latency=8190 ms total_cost=$0.07640 successes=5 • GPT-5.4 avg_latency=8783 ms total_cost=$0.03493 successes=5 Commentary: Across 5 real test queries, GPT-4o-mini was both cheapest ($0.0016 total) and fastest (5579 ms avg). Clear winner — 98% cheaper and 36% faster than the slowest pick. ``` --- ## What XFMS does for you Beyond ranking, XFMS gives you these levers to honor what you actually meant: - **`--primary `** — sacrosanct user preference. When you say *"cheapest model, period"*, the engine switches to lexicographic ranking: cost wins, other dimensions only break ties. No more weighted-blend surprises. - **`--ab`** — runs the top 3 picks against 5 real test queries (expanding to 10 or 15 if results trade off) and surfaces commentary on who won what. Grounds the recommendation in actual model behavior, not just benchmarks. - **`--strict-priorities`** — when you name two co-equal drivers ("cheap but high quality too"), the engine refuses to silently blend; it asks you which way to break the tie. - **Latent-requirement suggestions** — engine surfaces capabilities you didn't ask for but probably need (streaming for real-time chat, vision for OCR), so you don't get burned by what you didn't know. - **Deterministic by design** — every internal model call is content- cached; same input always returns the same answer. The "I got different picks for the same question" failure mode is gone. --- ## Install — add one URL to your AI client XFMS is a hosted Model Context Protocol (MCP) server. There is no package to install on your machine. You point your AI assistant — Claude Code, Cursor, Continue, Cline, or any MCP-speaking host — at the URL below and the tools appear inside your chat: **`https://xfms.xpansion.dev/mcp/`** That's it. The three discovery tools — `rank`, `pick`, `discover` — are free and work with no key. Your AI assistant does the small internal thinking work; we pay for nothing on your behalf, and you pay for nothing either. The two live-probe tools — `compare` and `benchmark` — actually run test queries against the real candidate models on OpenRouter. That inference cost rides with you, so they require your own OpenRouter key in an `X-OpenRouter-Key` header. You probably already have one — you're an OpenRouter model picker's target audience. If not, grab one at [openrouter.ai/keys](https://openrouter.ai/keys). The key travels encrypted to our server, is never logged, never persisted — used once per request and dropped. Same security posture as every other API key in your MCP config. Concrete install snippets for each AI client are in the next section. --- ## Use it inside Claude Code, Cursor, or any MCP client XFMS speaks **Model Context Protocol** (MCP) — the standard your AI assistant uses to call external tools. Once connected, you can ask the assistant *"which model should I use for OCR on shipping manifests?"* and it calls XFMS for you. No leaving the chat. No copy-pasting between windows. ### Hosted install — one line, no install required The XFMS engine hosts the MCP server itself at **`https://xfms.xpansion.dev/mcp/`**. Two install shapes depending on which tools you want. ### The free three — `rank`, `pick`, `discover` Just point your client at the URL. No key, no signup. **Claude Code:** ```bash claude mcp add xfms --transport http https://xfms.xpansion.dev/mcp/ ``` **Cursor** (`~/.cursor/mcp.json`) — or paste through *Settings → MCP*: ```json { "mcpServers": { "xfms": { "url": "https://xfms.xpansion.dev/mcp/" } } } ``` ### All five tools — adds `compare` and `benchmark` These two run real test queries against the actual candidate models on OpenRouter, so they require your OpenRouter key in an `X-OpenRouter-Key` header. Same install, one extra line: **Claude Code:** ```bash claude mcp add xfms --transport http https://xfms.xpansion.dev/mcp/ \ --header "X-OpenRouter-Key: sk-or-v1-your-key-here" ``` **Cursor** (`~/.cursor/mcp.json`): ```json { "mcpServers": { "xfms": { "url": "https://xfms.xpansion.dev/mcp/", "headers": { "X-OpenRouter-Key": "sk-or-v1-your-key-here" } } } } ``` **Continue / Cline / any other MCP host** — same URL + headers pattern; check your host's docs for the JSON config shape. Don't have an OpenRouter key yet? Grab one at [openrouter.ai/keys](https://openrouter.ai/keys). Restart your client, then ask it: > *"Use XFMS to pick a model for summarizing long legal contracts."* Five tools are available to the assistant: **`rank`** (a ranked shortlist), **`pick`** (the single best pick), **`discover`** (which quality dimensions matter for your purpose, without ranking), **`compare`** (live A/B between models you've already named), and **`benchmark`** (live A/B against the engine's top 3 picks). `compare` and `benchmark` require the `X-OpenRouter-Key` header above; the other three don't. --- ## Override the system's inference If you know which quality dimension matters most for your task, say so — your preference always wins over the LLM's inference: ```bash xfms rank "code refactor" --leaf-priorities "structured_output_reliability=1.0,factuality=0.5" ``` ```python xfms.rank( "code refactor", leaf_priorities={"structured_output_reliability": 1.0, "factuality": 0.5}, ) ``` --- ## How XFMS picks — the four principles Methodology in full at [`docs/methodology.md`](docs/methodology.md). The short version: 1. **No provider self-reports.** Every score comes from a third-party evaluator running the same protocol across every model. 2. **No single-source dependence.** Eight independent benchmark sources contribute today; no single leaderboard determines a pick. 3. **User intent beats LLM inference.** The system infers weights from your purpose, but your stated `leaf_priorities` always override the inference. 4. **Honest gaps over invented signal.** Missing data is recorded as missing — no interpolation, no synthetic scores. Coverage gaps surface on every pick. --- ## Part of the Xpansion Framework XFMS doesn't stand alone — it's the model-selection layer of the **[Xpansion Framework](https://xpansion.dev)**. ### The Xpansion thesis > **Humans communicate with intent compressed by contextual experience. > AI simply predicts patterns in language. Xpansion is the execution > layer that bridges them.** Every sentence a human types carries lifetimes of context that the speaker assumes the other side will decompress — what counts as "good enough," which constraints are non-negotiable, what failures last month taught them, what their house style demands. AI doesn't share that context. It predicts patterns in language, filling in the gaps with whatever's plausible to its training data. The result reads as plausible but isn't intent-honoring: sessions lose context, security holes ship silently, contracts break without warning, and there's no way to verify that what was built actually matches what was asked for. **They don't know what they don't know, and neither does AI.** Xpansion closes the gap. It **decompresses finite intent upfront**, **enforces it through code-driven AI behavior**, and **delivers binary-verified results against the intent across persistent memory** that survives every session boundary. ### Model Source — the model-selection enforcement When you say *"the best model for this task"*, you're compressing a lot: what counts as *best* depends on whether you care about factual reasoning or coherent prose, whether cost matters more than latency, whether you actually need vision or just text, whether the call has to stream, whether a particular benchmark dominates your real workload. AI on its own predicts the pattern — *what model do most people pick for queries that look like this?* — and gives you a plausible-sounding answer that's often wrong for *you*. XFMS does the decompression. It takes your stated purpose, infers which benchmarks actually map to it, honors your stated primary preferences without silently overriding them, surfaces the latent requirements you didn't know to ask about (streaming for real-time chat, vision for OCR), and probes the top picks against your real query to verify the recommendation — not predict it. Then it tells you, in plain English, why it picked what it picked. ### One module per enforcement The rest of the Xpansion stack enforces the same decompress- enforce-verify contract for different parts of the work: - **Dispatch** (`Dispatch`) — runtime task router. Watches what kind of work you're doing and routes it to the right tool. - **Finite Intent** (`XFFI`) — turns *"build me a feature"* into a finite spec with binary terminals *before* any code gets written. Stops scope drift at the source. - **Boundary Auditor** (`XFBA`) — checks every code edit against contracts. Stops broken function signatures and mismatched types from ever reaching production. - **Systemic Impact Analysis** (`XSIA`) — maps the blast radius of a proposed change before it lands. Tells you what else might break. - **Token Conservation** (`XFTC`) — manages how much of the conversation has to stay in the assistant's working memory. Prevents context loss in long sessions. - **Execution Audit** (`XFXA`) — verifies every promise from the spec was actually met before declaring a task done. The final binary check. - **Memory Tree** (`XFMT`) — session snapshots that stay searchable across conversations. Your assistant remembers what you decided last week. - **Security Auditor** (`XFSA`) — static + AI security scanning on every code edit. Catches secrets, injection paths, and unsafe patterns before they ship. The full picture, with the rest of the modules, lives at [xpansion.dev](https://xpansion.dev). **Xpansion is in pre-signup right now.** Early access and founding licenses are open at [xpansion.dev](https://xpansion.dev). XFMS is the first piece to ship public + free — the rest follow. --- ## Local development ```bash git clone https://github.com/VisionAIrySE/XFMS.git cd XFMS python3 -m venv .venv .venv/bin/pip install -e .[dev] .venv/bin/python -m pytest tests/ -v ``` The tests mock the HTTP layer so they run offline — no API keys needed to develop. --- ## License This client library is MIT-licensed. The recommender engine, the catalog, and the ingestion pipeline are not open source. See [`NOTICE`](NOTICE) for the patent reservation language and the relationship to the broader Xpansion Framework IP. --- ## Contact - **Russ Wright** — russ@visionairy.biz - **Xpansion Framework** — [xpansion.dev](https://xpansion.dev) - **Security disclosures** — see [`SECURITY.md`](SECURITY.md) ## Source & license This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [VisionAIrySE](https://github.com/VisionAIrySE) - **Source:** [VisionAIrySE/XFMS](https://github.com/VisionAIrySE/XFMS) - **License:** MIT - **Homepage:** https://xpansion.dev Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Versions - **0.4.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/mcp-visionairyse-xfms - Seller: https://agentstack.voostack.com/s/visionairyse - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.