MCP server for Dispatcharr IPTV stream management
Dispatcharr MCP demonstrates solid fundamentals: all 17 tools have clear verb-noun naming (list_, get_, create_, update_, delete_), descriptions range 50 - 200 chars (within baseline), and input schemas are fully visible with type definitions. However, output schemas are undocumented, responses are inferred from code but not formally declared. Parameter descriptions are present but terse (avg ~40 chars vs baseline 72). Error handling is absent: no recovery guidance, no categorization of retryable vs fatal errors. Destructive tools (delete_channel, delete_channel_group) lack confirmation patterns. Tool composition is sound (single responsibility, proper chaining via IDs), but pagination limits are not enforced in code, list_channels accepts page/page_size but no cap is visible, risking context window exhaustion.
Create a new channel. Provide at minimum a `name`. Optionally assign a channel number and group.
Create a channel directly from an existing stream. This is the quick way to add a channel: provide the `stream_id` and Dispatcharr will create a matching channel and link the stream. Optionally supply a custom `name`, `channel_number`, and which `channel_profile_ids` the new channel should belong to (omit for all profiles).
Create a new channel group.
Delete a channel by ID.
Delete a channel group by ID.
Get a single channel by its integer ID.
Output schemas undocumented. Tool descriptions state what is returned (e.g., 'Returns a paginated list') but formal response schemas are absent. LLMs cannot plan downstream tool calls or extract required fields without knowing response structure.
Destructive tools lack confirmation or dry-run patterns. delete_channel and delete_channel_group accept only an ID with no confirmation step. Agents can irreversibly delete resources without safeguards.
No error handling guidance. Tool descriptions do not explain what errors are possible, when to retry, or how to recover. A failed API call returns raw HTTP status without actionable next steps for the LLM.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 67 | 2026-07-28+ | v2 |
Get the live proxy status for a specific channel. `channel_id` is the channel's string identifier used by the proxy (typically the channel number or UUID).
Return a minimal stats delta for streams attached to a channel. Used to poll live stream health without fetching full stream objects. `ids` filters to specific stream IDs; `since` is an ISO-8601 timestamp for incremental updates.
Get all streams (M3U sources) assigned to a specific channel.
Get channels occupying channel numbers in a range. `start` is required. `end` defaults to `start` if omitted (single number lookup). Includes channels whose effective number is set via an override.
Get live status of all currently active proxy streams. Returns client counts, current stream URLs, buffer stats, and more for every channel that is actively streaming right now.
Get a single M3U stream by its integer ID.
List all channel groups.
List channels with optional filtering. Returns a paginated list of channels. Use `search` to filter by name, `channel_group` to filter by group name, and `page`/`page_size` for pagination.
List raw M3U streams from provider accounts. These are the source streams imported from M3U playlists, not the channel output. Filter by `m3u_account` (account ID), group name, or free-text `search`.
Preview regex find/replace results for streams in a channel group. Shows what names would be produced before committing a bulk regex rename. `channel_group` is required. `find` and `replace` define the substitution; `match` filters which stream names are affected; `exclude` skips matching names; `limit` caps the number of results returned.
Partially update a channel. Pass any subset of channel fields as `fields` (e.g. {"name": "BBC One", "channel_number": 1.0}). Only provided fields are changed.
Pagination limits not enforced. list_channels and list_streams accept page_size but no maximum is documented or enforced in code. An LLM could request 10,000 results, exhausting context window.
Parameter descriptions are terse (avg ~40 chars vs baseline 72). Many lack format/constraint details. E.g., 'Filter by channel name' does not explain case sensitivity, partial vs exact match, or regex support.