Agent quickstart

Integrate Emote into this project so our agent can react to incoming messages while its main reply runs independently.

Read https://useemote.com/docs/quickstart, https://useemote.com/docs/reference/request, and https://useemote.com/openapi.json first, then the relevant integration guide. Follow the linked references for custom reactions, responses, errors, and usage.

Inspect the message handler and agent instructions. Reuse existing reaction UI or the platform’s native reactions. If this is a custom chat without reactions, add a small accessible badge to the original incoming user message, matching the existing design. Ask me only if you cannot identify which chat interface or messaging platform to integrate.

Call POST https://useemote.com/v1/react from the server using EMOTE_API_KEY. Send the incoming message, the agent’s personality, recent conversation context including previous reactions, and supported reaction options. Use built-in emoji strings or custom {id, description} objects. Respect the documented input limits. Load the key in the server environment and restart after configuration. Never put it in client code or ask me to paste it into chat. If the key is missing, finish the integration and explain the remaining setup.

Run this alongside the main model, not as a model tool call. Parse the response as a JSON string. If it is "none", do nothing; otherwise map the returned emoji or custom ID to our platform’s reaction and attach it to the original message ID. Use a stable Idempotency-Key per message and preserve the same request body on retries. Deduplicate incoming events and reaction delivery separately.

Handle timeouts and API failures without interrupting replies. Ignore deleted messages and stale results. Add focused checks for an emoji, "none", and API failure. With a configured key, test a fresh incoming message in the actual chat and verify any chosen reaction appears on that message while the normal reply still works. Distinguish live API verification from mocked tests. Follow this project’s conventions and report what changed and any setup still needed.