Build with agents
Add video uploads, playback, and analytics to your application with a coding agent.
The Chunkify skill guides your coding agent through adding video uploads, streaming, playback, and analytics. It uses Chunkify's MCP connection, SDKs, components, and documentation.
The agent inspects your application and follows its existing framework, data models, and UI conventions. You can also use the skill when starting a new application.
Install the skill
Open your application in your coding agent and use the installation prompt for that agent:
Ask Codex to install the skill in your project:
Install the Chunkify skill from https://chunkify.dev/SKILL.md in .agents/skills/chunkify/ in this project. Include its linked reference files and keep their references/ directory structure.
Mention $chunkify in your request to use the installed skill. See Codex skills for other installation locations.
Both agents use the same skill files:
chunkify/
├── SKILL.md
└── references/
├── asset-flow.md
├── cors.md
├── mcp.md
└── migrate-jobs.md
Other coding agents that support Agent Skills may use a different installation directory. Follow your agent's skill setup instructions.
Set up your Project
Connect your agent to https://api.chunkify.dev/mcp using OAuth. Sign in and select the Projects it can access.
The skill uses this connection to inspect resources, configure hosted or self-hosted analytics connections, and query video performance.
Your application's server needs a Project API token for uploads and Asset creation. The agent can create one through MCP and save it privately as CHUNKIFY_PROJECT_TOKEN, or reuse a working credential.
Webhook handlers also need CHUNKIFY_WEBHOOK_SECRET from the webhook configuration.
MCP can create, list, and revoke Project and Team tokens. Team token management requires separate consent because those tokens cover every Project in the team, including unselected and future Projects.
Use a Project token for media integration. Managing analytics through OAuth needs no Team token.
Keep tokens and webhook secrets out of chat, tracked files, and browser code. Tokens created through MCP keep working after you disconnect the assistant. Revoke them separately to stop applications that use them.
If MCP is unavailable, the skill can use the Chunkify CLI or official SDK with authorized credentials. CLI configuration is separate from the application's server environment. Tell the agent if you use a named CLI profile.
For ongoing playback, connect your storage and configure CDN delivery. The skill uses your Project's configured storage and CDN unless you request another setup.
Playback from existing URLs and standalone playback analytics do not require a Project token.
Describe what you want to build
The following prompts work with either agent. Explain the flow you need and any application requirements. For example:
Use the Chunkify skill to add video uploads and on-demand playback to this application. Use my Project's configured storage and CDN, and keep the existing UI and authorization. Handle playback readiness through verified webhooks and save playback URLs with our video records.
You can also ask for a smaller integration, such as analytics for an existing player:
Use the Chunkify skill to add playback analytics to our existing player. Send events to our /api/video-events endpoint and use our video record ID to identify each video.
See Chunkify Player and playback analytics for the available controls and options.
Verify the result
Ask the agent to run your application's checks and test the complete flow. For uploads, verify that a video reaches playback and that webhook updates reach your application's video record.
For analytics, verify that a playback session reaches your chosen destination. The agent should report any setup or checks that remain incomplete.