Fishjam Platform

Q: Why use Fishjam Platform on TypingMind?

Because you install it once and use it with any model. Fishjam Platform is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

Q: How do I install Fishjam Platform in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/software-mansion-labs/skills/tree/main/skills/fishjam/references/platform. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Q: Which AI models can use Fishjam Platform?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

Q: How many AI models can I use with Fishjam Platform?

As many as you like. As long as a model supports skills, you can use Fishjam Platform with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Q: Is the Fishjam Platform AI skill free?

Yes. It is published on GitHub by software-mansion-labs under the MIT license. You only pay your own AI provider for the tokens you use.

Q: What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

Q: How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

Organization

software-mansion-labs

fishjam-platform

Fishjam platform fundamentals — domain model and auth shared by all SDKs. Covers glossary (room, peer, track, agent, streamer, viewer), the four room types (conference / audio_only / livestream / audio_only_livestream), two-tier auth (management vs peer tokens), Sandbox vs production, WebSocket-vs-webhook notifications and the 18-event taxonomy, REST endpoint list, and the end-to-end auth/join flow. Read this first for any Fishjam task. Trigger on: 'Fishjam', 'fishjam.io', 'fishjam.swmansion.com', 'fishjam dashboard', 'management token', 'peer token', 'sandbox API', 'useSandbox', 'room manager', 'livestream room', 'audio-only room', 'conference room', 'Fishjam ID', 'room types', 'auth flow', 'webhook', 'webhookUrl', 'FishjamWSNotifier', 'FishjamNotifier', 'notifier vs webhook', 'ServerMessage', 'peerAdded', 'peerConnected', 'roomCreated', 'refreshPeerToken', 'rest endpoint', 'OpenAPI', 'llms.txt'.

Overview

Publisher	software-mansion-labs
Repository	`skills`
Skill name	`fishjam-platform`
Stars	296
Forks	21
Bundled files	9
License	MIT
Links	Skill source Repository

Markdown instructions
A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.
Works with any LLM
AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.
9 bundled files
Scripts, templates, and references the model can read while it works. Files are read-only and never executed.
Open source
Published by software-mansion-labs on GitHub. Read the source before you install it.

Installation

Install the Fishjam Platform AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

Open the app and go to Plugins → Skills.
Choose "Install from GitHub".
Paste the skill folder URL below and confirm.
https://github.com/software-mansion-labs/skills/tree/main/skills/fishjam/references/platform
Enable the skill in any chat where you want it available.

Store

Installed

Skills

Add skills

Import one or more skills from a public GitHub folder or a local .skill or .zip archive, or SKILL.md file.

From GitHub URLFrom .skill, .zip, or SKILL.md

GitHub repo or folder

https://github.com/software-mansion-labs/skills/tree/main/skills/fishjam/references/platform

Points to a single skill folder, or any folder that contains multiple skills. Public repos only.

CancelContinue

Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills

git clone --depth 1 https://github.com/software-mansion-labs/skills.git /tmp/skills
mkdir -p .claude/skills
cp -r /tmp/skills/skills/fishjam/references/platform .claude/skills/fishjam-platform

Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Fishjam Platform in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Fishjam Platform on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Fishjam Platform is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

Loaded only when it is needed
The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.
Switch models mid-chat
Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Fishjam Platform

Software Mansion's hosted WebRTC streaming toolkit. Fishjam takes care of media servers, peer networking, and codec negotiation so applications can build video conferencing, audio chat, and one-to-many livestreaming with sub-second latency.

This sub-skill defines the domain model, auth, room types, and notifications that every SDK builds on. Read it first; then drop into the SDK sub-skill you need.

A production Fishjam app always has a backend (server SDK) + a client (client SDK). The backend mints peer tokens; the client uses them to connect. For prototyping you can skip the backend with the Sandbox API — see sandbox-vs-production.md.

Key concepts at a glance

Room — a collection of peers exchanging tracks. Three primary room types: conference (default, bidirectional), audio_only (75% cheaper, video silently dropped), livestream (one streamer, many viewers, <300 ms). audio_only_livestream combines the audio-only and livestream discounts.
Peer — a participant in a room. Can be a webrtc client, an agent (server-side program), or a vapi voice agent.
Track — one audio or video stream from a peer (camera, mic, screen share, etc.). One peer can publish multiple tracks.
Streamer / Viewer — livestream-room-specific peer roles (one streamer, many viewers).
Fishjam ID — your instance identifier; goes in both backend and client config.
Management Token — long-lived admin secret; backend only.
Peer Token — 24h, scoped to one peer in one room; safe for clients.

Full definitions: glossary.md.

Room types

Type	Direction	Cost	Use case
`conference`	many ↔ many	base	Meetings, classrooms, interactive webinars
`audio_only`	many ↔ many (audio only)	75% off	Voice rooms, podcasts, town halls
`livestream`	one → many	20% off	Live events, broadcasts, sports

audio_only_livestream is 75% off the livestream rate (equivalently, 80% off conference) — the cheapest mode for one-to-many audio.

Detail + decision matrix: room-types.md.

Auth flow (5-step sketch)

Backend → Fishjam: createPeer(roomId) authenticated with the management token.
Fishjam → Backend: returns { peer, peerToken }.
Backend → Client: ships peerToken to the frontend (never the management token).
Client → Fishjam: joinRoom({ peerToken }).
Fishjam: completes the WebRTC handshake; the peer is now in the room.

Full diagram + token-refresh + livestream + MoQ tokens: auth-model.md.

Sandbox vs production

Sandbox API — no-auth HTTP backend hosted by Fishjam. For prototyping only. Generated URL must never be shipped in a production build (anyone with it can create rooms on your behalf).
Production — your own backend running a server SDK. Required for any real-world deployment.

Migration recipe: sandbox-vs-production.md.

Notifications: WebSocket or webhook?

Both deliver the same 18-event set surfaced by the SDKs (room/peer/track/streamer/viewer/channel lifecycle — the underlying protobuf ServerMessage carries additional oneof fields the SDKs intentionally ignore). Pick by deployment shape:

Long-lived worker, single process → FishjamWSNotifier / FishjamNotifier (persistent socket).
Serverless, scaled out, behind a load balancer → webhooks (set webhookUrl per room, decode binary protobuf).

Detail: notifier-vs-webhook.md and notifications-taxonomy.md.

Key rules

Management token never leaves the backend. If it leaks, regenerate it.
Peer tokens expire 24 hours after creation. Refresh via refreshPeerToken / refresh_peer_token if you need longer.
Sandbox API URL is not a secret in the cryptographic sense, but anyone holding it can create rooms — treat it like a dev credential and rotate from the Dashboard if exposed.
Conference is the default room type. Use audio_only for voice apps (much cheaper). Use livestream only when one peer broadcasts and others only watch.
Livestream rooms accept exactly one video track + one audio track. Additional tracks are silently dropped.

References

File	When to read
`glossary.md`	Define a term (room, peer, track, agent, etc.) precisely.
`room-types.md`	Pick `conference` vs `audio_only` vs `livestream`. Cost matrix. Codec defaults.
`auth-model.md`	Token types, lifetimes, refresh, livestream / MoQ tokens, full auth flow.
`sandbox-vs-production.md`	Use the Sandbox API for dev; migrate to a real backend.
`notifier-vs-webhook.md`	Pick WebSocket notifier or webhook receiver.
`notifications-taxonomy.md`	All 18 SDK-surfaced server notification events with payload shapes.
`rest-endpoints.md`	Raw HTTP surface — flat endpoint list, auth header, deprecation header.
`lifecycle-flow.md`	End-to-end client↔backend↔Fishjam sequence.
`llms-and-docs.md`	Pointers to upstream docs, llms.txt, OpenAPI spec, protobuf, dashboard.

Mixing several streams into one composed output (a call as a broadcast, restreaming, overlays) is covered in ../composition/SKILL.md.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Fishjam Platform AI skill do?

Why use Fishjam Platform on TypingMind?

Because you install it once and use it with any model. Fishjam Platform is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Fishjam Platform in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/software-mansion-labs/skills/tree/main/skills/fishjam/references/platform. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Fishjam Platform?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Fishjam Platform?

As many as you like. As long as a model supports skills, you can use Fishjam Platform with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Fishjam Platform AI skill free?

Yes. It is published on GitHub by software-mansion-labs under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Banner Design

nextlevelbuilder

OrganizationPopular

Design banners for social media, ads, website heroes, creative assets, and print. Multiple art direction options with optional generated or supplied visuals. Actions: design, create, generate banner.…

Distributed Triage

Sub-triages issues in the oncall:distributed queue by assigning distributed module labels, routing to sub-oncalls, and marking triaged. Use when an issue has been routed to oncall:distributed and…

Hyperframes

Create video compositions, animations, title cards, overlays, captions, voiceovers, audio-reactive visuals, and scene transitions in HyperFrames HTML. Use when asked to build any HTML-based video…

Audio Jingle

Audio generation skill — jingles, beds, voiceover, and sound effects. Routes music requests to Suno V5 / Udio / Lyria, speech to MiniMax TTS / FishAudio / ElevenLabs V3, and SFX to ElevenLabs SFX or…

Overview

Installation

Install in TypingMind

Install in another agent

Use it in TypingMind

Works with any AI model

Skill instructions

Fishjam Platform

Key concepts at a glance

Room types

Auth flow (5-step sketch)

Sandbox vs production

Notifications: WebSocket or webhook?

Key rules

References

Bundled files

Frequently asked questions

Related AI Skills

Set up your own AI workspace now