MCP server wrapping Agnes Video 2.5 / 2.5 Flash with reference-mode support (mode=text|keyframe|reference, images/audio/video references)
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"agnes-video-25-mcp","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# Agnes Video 2.5 MCP Service Standalone MCP server wrapping **Agnes Video 2.5**, **Agnes Video 2.5 Flash**, and **Agnes Video v2.0** (whitelist, since v0.2.0). First-class `reference` mode support (image / audio / video reference via `images[]` / `audios[]` / `videos[]` arrays and `<Picture N>` / <Audio N>` / `<Video N>` prompt placeholders) is available on 2.5 / 2.5 Flash; v2.0 uses its own protocol (`ti2vid` / `keyframes` + `extra_body.image[]`) which the server translates transparently — callers keep the unified input shape. This service is **independent** of `hosted_mcps/agnes-mcp-studio/` (which targets the legacy `agnes-video-v2.0` model via the older `ti2vid` / `keyframes` API). Run it alongside or replace — both can coexist. ## When to use this service | Need | Use this service | Use `agnes-mcp-studio` | |---|---|---| | Reference-mode video (images / audios as visual + audio anchors) | ✅ (2.5 / 2.5 Flash) | ❌ (only `extra_body.image` keyframes) | | First / last frame control (`first_frame` / `last_frame`) | ✅ (2.5 / 2.5 Flash / v2.0) | ✅ (raw `ti2vid` / `keyframes`) | | Legacy v2.0 model (back-compat with existing pipelines) | ✅ (0.2.0+ whitelist; unified inputs) | ✅ (raw v2.0 protocol) | | Image generation (text-to-image, image edit, 2.5-flash) | ✅ | ✅ (parallel surface) | ## Install ```bash cd hosted_mcps/agnes-video-25 uv sync # installs the package + dependencies ``` ## Run standalone (stdio MCP) ```bash uv run agnes-video-25-mcp ``` ## Wire into `.mcp.json` Add alongside `multimedia-creator`: ```jsonc { "mcpServers": { "multimedia-creator": { /* existing agnes-mcp-studio entry */ }, "agnes-video-25": { "command": "uv", "args": [ "--directory", "/absolute/path/to/hosted_mcps/agnes-video-25", "run", "agnes-video-25-mcp" ], "env": { "AGNES_API_KEY": "sk-your-key-here", "AGNES_BASE_URL": "https://api.agnes-ai.cn/v1" } } } } ``` ## Environment variables…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.