Gemini MCP Server

Language models generate; Concordance verifies. The verify tool checks a claim deterministically (no model in the loop) and returns HOLDS / BROKEN / INCOMPLETE with the worked reasoning and a sealed receipt (content_hash + cite_url) that re-fetches byte-identical or not at all. Also: ranked search over an ~11k-record library, seal_fetch to re-verify any receipt, redact to strip PII before text travels, and a sealed connection graph. Runs sovereign/offline too (stdlib-first Python). A public false-positive benchmark covers every domain: the engine has never sealed a falsehood. 38 tools live; remote endpoint at https://narrowhighway.com/mcp.

Is this song AI or human? SongCheck detects AI-generated music (Suno, Udio, and more) and media from any AI agent. Point it at an audio file, an image/video, or a whole music folder and it returns a verdict (LIKELY AI-GENERATED / UNCERTAIN / LIKELY HUMAN), an AI-probability score, confidence, and provenance signals (Content Credentials / SynthID watermark, generator hints). Tools: detect_ai_music, detect_ai_media, scan_catalog (audit an entire catalog), and songcheck_health. Free tier is 5 checks/day; a paid key unlocks unlimited catalog scans. Powered by SongCheck (Khaled Media), a self-hosted v9 ensemble detector.

Official remote MCP server for Framesail AI. Create long-form (faceless YouTube) videos end to end from any MCP client: script, locked character references, storyboard, voiceover, and final video editing — with characters and style held consistent across every shot. Making long-form AI video today means 8+ tabs stitched by hand — an LLM for the script, a voice model, an image model, a video model — with characters drifting between tools and style resetting at every export. Framesail replaces the patchwork: the whole pipeline runs in one place and manages your video's context end to end. Six stages: Style (paste images, videos, or YouTube links and Framesail reverse-engineers the look, voice, and direction), Script (write it yourself or generate it in your narrative style), Reference images (auto-generated for every character, place, and prop), Voiceover (one narrator or many characters, with word-level timing), Storyboard (planned scene by scene), and Editor (captions, music, SFX, then export). No black box: you control every prompt, asset, model, and setting.

Query verified 2026 US dental cost data from Real Dental Costs: average implant, veneer, braces and procedure prices for all 50 states and 206 cities, state-vs-state comparisons, and the composite US Dental Cost Index. Open data (CC BY 4.0, Zenodo DOIs). Every response includes the source and a reference URL. Pricing/market research, not medical advice. Data: https://realdentalcosts.com

Public, read-only MCP over the complete 64-hexagram I Ching corpus — hexagrams, trigrams, changing lines, and reading context from an original English translation. Deterministic lookup only; no casting or divination.

ContextBridge (CB) is a local-first retrieval layer that sits between your codebase and your AI coding agent. Instead of pasting raw source files into context or letting the AI guess which files matter, CB indexes your codebase's real structure (via Graphify) and returns a compact, ranked result — owner file, related files, key symbols, and dependency chains — grounded in your actual code. A typical response is ~4–8 KB instead of the 100+ KB of raw source an AI would otherwise need to read to answer the same question — roughly a 96% cut in input tokens sent to your cloud AI, with zero hallucinated file paths or method names.

VC follow intelligence for AI agents. Track what top investors follow on X — detect new follows, convergence signals, and trending companies before they're announced.

5 AI skills as MCP tools: MBTI relationship sandbox (2-20 people), fact-checking (3 methods: source-tracing, cross-reference, logic-check), deep business analysis (5-layer), game audio brief generator, weekly skill radar. All monetized via Alipay AI Pay (¥0.10-0.50/call).