# Openrouter Streaming Setup

> Implement streaming responses with OpenRouter. Use when building real-time chat interfaces or reducing time-to-first-token. Trigger with phrases like 'openrouter streaming', 'openrouter sse', 'stream response', 'real-time openrouter'.

- Skill: `micsapp/openrouter-streaming-setup` (Agent Skill, multi-file: 9 files)
- Install (CLI): `npx skillmds@latest add micsapp/openrouter-streaming-setup`
- Raw SKILL.md: https://api.skillmd.com/api/skills/micsapp/openrouter-streaming-setup/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- License: MIT
- Author: micsapp (https://skillmd.com/u/micsapp)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/micsapp/openrouter-streaming-setup

---

# Openrouter Streaming Setup

## Overview

This skill demonstrates streaming response implementation for lower perceived latency and real-time output display.

## Prerequisites

- OpenRouter integration
- Frontend capable of handling SSE/streaming

## Instructions

1. **Enable streaming**: Set `stream: true` in your chat completion request body
2. **Handle SSE chunks**: Parse each `data: {...}` line from the response stream, extracting `choices[0].delta.content` from each chunk
3. **Detect stream end**: Watch for `data: [DONE]` to know when the stream is complete; accumulate chunks for the full response
4. **Implement frontend rendering**: Use `ReadableStream` or `EventSource` in the browser to display tokens as they arrive
5. **Add error recovery**: Handle mid-stream disconnections with automatic retry and partial response preservation

## Output

- Real-time token-by-token text output in the UI
- Reduced time-to-first-token compared to non-streaming requests
- Complete response assembled from all chunks with usage stats from the final chunk

## Error Handling

| Error | Cause | Fix |
|-------|-------|-----|
| Stream cuts off mid-response | Network timeout or model error | Implement reconnection logic; save partial output |
| Missing `usage` in stream | Some models omit usage in streaming mode | Set `stream_options: { include_usage: true }` or make a separate token count call |
| Empty delta chunks | Keep-alive pings from the server | Filter chunks where `delta.content` is null or empty |

See `${CLAUDE_SKILL_DIR}/references/errors.md` for full error reference.

## Examples

See `${CLAUDE_SKILL_DIR}/references/examples.md` for runnable code samples.

## Resources

- [OpenRouter Documentation](https://openrouter.ai/docs)
- [OpenRouter Models](https://openrouter.ai/models)
- [OpenRouter API Reference](https://openrouter.ai/docs/api-reference)
- [OpenRouter Status](https://status.openrouter.ai)