`docs/CONTRACT-mcp.md` goes to mcp-v2 with a change table at the top, because v1 lasted a day and an agent that learned it deserves to be told what moved rather than to discover it. The new §4 freezes the resource surface: the eight URIs, their bodies, their query parameters, and the five rules the layer keeps (a read never writes; only declared parameters; a literal route beats a template; a frame's media type comes from its bytes; no subscriptions). §6 gains the other half of the error mapping -- for a resource read every exit code arrives as `-32602` with the remedy in `message` and the structured object in `data`, because `ReadResourceResult` has no `isError` to use instead. §2 records what `--toolsets` is and is not: there is no grouping primitive in MCP, the `Tool` object has nowhere to put a family, and tools are a flat list in the specification too -- so the only lever is which tools the server lists, decided by whoever spawns it. README.md leads with the split rather than with a tool count, since that is the thing a reader has to understand before any of the names make sense, and shows the `--toolsets` form for a client with a tight budget. ARCHITECTURE.md's sixth box grows its second half and says why `shot` is on both sides. A second TODO, `mcp-resources-cannot-be-subscribed-to.md`, records the limit that cost the most: `Server.get_capabilities` hardcodes `ResourcesCapability(subscribe=False, …)`, so registering a `subscribe_resource` handler changes nothing and a client is correctly told not to try. Subscriptions would have been the principled answer to the missing `logs --follow` -- it is the one place this design wanted a capability the SDK will not advertise -- so the file names the one-line local override, says what it risks, and recommends upstreaming the real fix. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011ZeH65zd6pFTgBSu52W9PE
hyprland-headless-stream-plugin
A Nix flake that runs a headless Hyprland session and publishes its frames on a persistent Unix socket, so any process that can open a socket can see the screen without owning a display.
Two parts:
- NixOS module —
services.hyprland-headless-stream: starts Hyprland with no physical monitor, and a small frame server beside it. - Package —
hhstream— the frame server (publishes frames) and its client (reads them). Pure Python standard library.
{
inputs.hhstream.url = "git+https://<host>/git/<owner>/hyprland-headless-stream-plugin.git";
# in your host config
imports = [ inputs.hhstream.nixosModules.default ];
services.hyprland-headless-stream = {
enable = true;
# Optional: the default is one 1280x720@30 rule with an empty name, which
# is Hyprland's wildcard and applies to whatever outputs the GPU has.
outputs = [ { width = 1920; height = 1080; refresh = 60; } ];
};
}
Then any consumer:
hhstream-client --socket /run/hypr-stream/stream.sock --count 1 --out frame.png
The socket is a plain file descriptor: the server serves every client that connects, and keeps running when one disconnects.
Driving it from an agent: the MCP server
hhstream-mcp is the hhstream CLI behind a Model Context
Protocol surface, so an agent can look at this
desktop, put an application on it, and aim its one pointer and keyboard at that
application without writing command lines. The interface is frozen in
docs/CONTRACT-mcp.md.
It is split the way the protocol splits things:
- Resources are the state — eight of them, read by URI, and reading one
changes nothing.
hhstream://statusis whether the desktop is up,hhstream://sessionsis what is running,hhstream://sessions/<id>/logsis an application's own output,hhstream://pointeris who holds the input lease,hhstream://screenis a frame of the screen. - Tools are the acts — twenty-one, one per CLI verb that does something.
session_startputs an application on the screen,pointer_acquiretakes the input lease,keyboard_typetypes,session_killends a session.
A host decides what state to put in front of its model; the model decides what
to do. That is the whole of the split, and it is why shot is still a tool
(capturing is an act at a moment) while hhstream://screen exists beside it.
It is a thin wrapper and nothing more: every tool and every resource builds the argv list the CLI's own parser takes, runs the CLI's own handler, and maps what came back. The idle clock, the pointer lease, the allowlist and the exit-code table all stay where they were.
Install it with the module — off by default, because it is the surface through which a language model starts processes on this machine:
services.hyprland-headless-stream = {
enable = true;
sessions.enable = true; # the session, pointer and keyboard tools need the broker
mcp.enable = true; # puts `hhstream-mcp` on the system PATH
};
Then register it with a client. The transport is stdio — one process per client, spawned by the client, no port and no daemon — so the whole configuration is a command and an environment:
{
"mcpServers": {
"hhstream": {
"command": "hhstream-mcp",
"env": {
"HHSTREAM_SOCKET": "/run/hhstream/control.sock",
"HHSTREAM_FRAME_SOCKET": "/run/hypr-stream/stream.sock"
}
}
}
}
Both variables are the CLI's own and both are shown at their defaults, so they
can be left out unless this machine moved a socket. The account the client runs
as needs what the CLI needs: membership of the compositor's group (hhstream)
to read the frame socket, and of the broker's to reach the control socket. The
server cannot recover that for itself the way the CLI does — see
docs/CONTRACT-mcp.md §5 — so if shot reports it cannot read the socket, the
answer is "command": "sg", "args": ["hhstream", "-c", "hhstream-mcp"] or an
account in the group.
hhstream-mcp --list-tools and --list-resources print the two tables and
exit, which is the quickest way to check an install.
If twenty-one tools is more than a client has room for, --toolsets narrows
what is offered to whichever families are wanted — frames, sessions,
pointer, keyboard:
{ "command": "hhstream-mcp", "args": ["--toolsets", "frames,pointer,keyboard"] }
Resources are deliberately not narrowed by it: they are the host's context, not the model's tool budget. Nothing else about the server is configurable — it holds no state and caches no lease, and the pointer lease is an explicit argument on every gesture tool for exactly that reason.
| Document | What it is |
|---|---|
docs/CONTRACT.md |
the wire protocol and the module surface — the interface other people build against |
docs/CONTRACT-sessions.md |
the broker, the containers and the input lease |
docs/CONTRACT-mcp.md |
the MCP surface: the resource URIs, the tool names, their arguments and the error mapping |
docs/ARCHITECTURE.md |
how the pieces fit: compositor, capture, server, client |
README.dev.md |
how to build, test and hack on it |
examples/README.md |
the demo that films the screen: agent-eyes-demo.py |
Status: work in progress — the module and the test suite are being built.