What is plugsky-fusion?
When you set model="plugsky-fusion", Plugsky inspects each request and routes it to the optimal model in your tier based on:
- Token count (input + expected output)
- Tool call presence and complexity
- Prompt heuristics (domain, language, length)
- Per-workspace strategy setting
Strategies
Set the strategy per-workspace, per-API-key, or per-request:
bash
curl https://api.plugsky.com/v1/chat/completions \
-H "Authorization: Bearer $PLUGSKY_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "plugsky-fusion",
"fusion": {"strategy": "cost_saver"},
"messages": [{"role": "user", "content": "Hello!"}]
}'Strategy options
| Strategy | Behavior |
|---|---|
cost_saver | Cheapest model that can handle the call. Falls back to stronger on failure. |
balanced | Default. Mid-tier for most calls, cheap for trivial, strong for hard. |
max_quality | Always the strongest model in your tier. |
custom | Use your custom rules from the dashboard. |
Custom rules
On the dashboard, define rules like:
- Prompts with the word "refactor" → plugsky-pro
- Prompts with > 4000 input tokens → plugsky-frontier
- Prompts from free-plan users → plugsky-micro
- Prompts during business hours (9-5 GMT) → plugsky-frontier
Rules are evaluated in order; first match wins.
Frequently asked questions
Does routing add latency?
No. The router runs in the API gateway, adding <5ms to the request.
Can I override routing per request?
Yes. Use a specific model name like plugsky-pro to bypass routing for that request.
Can I see routing decisions in logs?
Yes. Every request log includes the chosen model, the strategy, and the rule that fired (for Custom).
Can I set the strategy per API key?
Yes. Per-workspace, per-API-key, or per-request. Default is the workspace setting.
Try plugsky-fusion
Set model="plugsky-fusion" and Plugsky picks the right model per call.