Media Router
Pick the right PostPlus media skill chain (transcription, subtitles, beats, B-roll, edit plans) before spending tokens on the wrong workflow.
Overview
Media Router is an agent skill most often used in Build (also Launch content, Grow content) that routes audio, video, and transcript jobs to the correct PostPlus media-understanding workflow before execution.
Install
npx skills add https://github.com/postplusai/postplus-skills --skill media-routerWhat is this skill?
- Routes by five dimensions: inputType, goal, quality, scale, and costMode—not file extension alone
- Maps audio→transcript, video, subtitles, semantic analysis, beats, broll-map, and edit-plan intents to downstream skills
- Enforces cheap-first vs quality-first model chains when rough output is sufficient
- Explicitly separates media understanding from generation-router (images/video generation)
- Does not execute collection or editing—only chooses the right workflow before execution
- Five routing dimensions: inputType, goal, quality, scale, costMode
- Six goal types including transcript, subtitles, semantic-analysis, beats, broll-map, and edit-plan
Adoption & trust: 1 installs on skills.sh; 8 GitHub stars; 2/3 security scanners passed (skills.sh audits); trending (+100% hot-view momentum).
What problem does it solve?
You have media files but the agent jumps straight to transcription or an expensive model chain without matching quality, scale, or edit intent.
Who is it for?
Solo creators using PostPlus skills who need consistent dispatch from mixed audio/video/transcript inputs to transcript, subtitle, semantic, or edit-plan workflows.
Skip if: Image or generative media requests (use generation-router), or teams that want the router to perform transcription or editing itself.
When should I use this skill?
User has audio, video, transcript, subtitle, A-roll, or B-roll material and the first decision is which skill and model chain should run for transcription, subtitles, beats, B-roll, or edit-ready outputs.
What do I get? / Deliverables
You get an explicit routing decision (which skill and quality/cost path to run) before any transcription, subtitle, or edit-prep skill executes.
- Routing decision naming target skill(s) and quality/cost path
- Structured dimensions: inputType, goal, quality, scale, costMode
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Media pipelines are wired during product and content build; this router sits at the integration layer that hands off to specialized skills. It is a routing contract between agent skills—not execution—so integrations is the canonical shelf for workflow dispatch.
Where it fits
Wire your agent so uploaded interview audio picks audio-transcription only when rough text is enough, not a full edit-plan chain.
Choose subtitle-ready routing before exporting captioned clips for social distribution.
Batch-route a backlog of videos by quality tier and costMode before semantic analysis or B-roll mapping.
How it compares
Use as a pre-flight router contract—not a replacement for audio-transcription or downstream edit skills.
Common Questions / FAQ
Who is media-router for?
Indie builders and content-focused agents using the PostPlus skill family who need the right media-understanding chain chosen before work starts.
When should I use media-router?
Use it in Build when wiring agent media pipelines, at Launch when preparing edit-ready or subtitle assets for distribution, and in Grow when batch-processing content libraries—whenever the first question is transcribe vs subtitles vs beats vs B-roll vs edit-plan.
Is media-router safe to install?
Review the Security Audits panel on this Prism page for install risk and upstream repo signals before enabling it in production agent workflows.
Workflow Chain
Then invoke: audio transcription
SKILL.md
READMESKILL.md - Media Router
# Media Router ## Use When - The user has audio, video, transcript, subtitle, A-roll, or B-roll material. - The first decision is whether to transcribe, package subtitles, add semantic understanding, or build an edit-ready plan. - The output quality target is rough, subtitle-ready, or edit-ready. ## Do Not Use When - Do not treat every media request as plain transcription. - Do not execute downstream collection or editing work in this router. - Do not choose a costly model chain when rough output is enough. - Do not route image, video, audio, or workflow generation here; use `generation-router`. ## Core Rule Route by job intent, not by input file type alone. Choose from five dimensions: - `inputType`: audio, video, transcript, or transcript+assets - `goal`: transcript, subtitles, semantic-analysis, beats, broll-map, or edit-plan - `quality`: rough, subtitle-ready, or edit-ready - `scale`: single or batch - `costMode`: cheap-first or quality-first ## Default Routing Logic - Audio to transcript -> `audio-transcription`. - Video to transcript or subtitles -> `video-transcription`. - Existing transcript to subtitle files -> `subtitle-packager`. - Transcript plus assets to cut logic or B-roll plan -> `editing-decision-engine`. - Video to edit-ready plan -> transcription first, optional subtitle packaging, optional semantic video analysis, then `editing-decision-engine`. - Analyze first, then generate -> this router handles the understanding step, then hands back to `generation-router`. ## Quality Thresholds - `rough`: quick review, search, logging, or candidate beat spotting. Prefer cheaper transcription and fewer enrichments. - `subtitle-ready`: captions, SRT/VTT, and reliable sentence or word timing. Timestamps are mandatory. - `edit-ready`: A-roll/B-roll decisions, cutaway placement, and beat-level planning. Use timestamped transcript plus semantic understanding when visual proof matters. ## Output Shape The artifact contains: - `route` and `why` - `primarySkill` - `supportingSkills` - `modelPlan` - `outputArtifacts` - `needsTimestamps` - `executionMode` ## Stop Conditions - Stop when required user intent, source evidence, or owned input artifacts are missing and guessing would change the result. - If an owned CLI or script command fails, report the exact error and stop. Do not bypass the failure with metadata-only answers, readiness probing, local payload rewrites, fallback providers, or unpublished tools. ## Handoff - Hand the route JSON to the named `primarySkill`. - Preserve expected artifacts so downstream skills know whether to produce `transcript.json`, `timed-transcript.json`, subtitles, beat maps, or edit plans. ## Public Command Boundary - Choose the smallest matching command or workflow from the user input and run it directly. - If an owned CLI or script command fails, report the exact error and stop. Do not bypass the failure with metadata-only answers, readiness probing, local payload rewrites, fallback providers, or unpublished tools. - This public skill is instruction-driven. Produce the artifact described by the workflow directly from the available evidence. - Do not call private provider/runtime paths or unpublished local tools. - If the CLI returns a quote-confirmation challenge, run `postplus quote confirm --json --challenge-file <challenge.json>` and retry with the returned token.