zoo-code-modelis-config

Drop-in provider config for Zoo Code / Roo Code / Cline / Continue — point at Modelis (OpenAI-compatible gateway, one key for ~95 models; modelis-auto billed per call within tier limits) in under 2 minutes.

Migrating off Roo Code (shut down)? This is the fastest path to a working provider config on the fork.

---

1. Get a key

New accounts get $1 free credit. Sign up at modelishub.com and copy your API key.

Add credit from the Modelis console when you need more:

• Card, via Gumroad
• Alipay or WeChat, via Afdian

Both are on https://modelishub.com/pricing.

2. Paste the config

Modelis is an OpenAI-compatible gateway. Base URL: https://modelishub.com/v1.

Zoo Code / Roo Code / Cline

Settings → API Provider: OpenAI Compatible

Field	Value
Base URL	https://modelishub.com/v1
API Key	your Modelis API key
Model ID	modelis-auto

modelis-auto is the standard tier (same as modelis-auto:standard). Or copy cline-settings.json values directly.

Continue

Copy continue-config.yaml into ~/.continue/config.yaml (or merge the models: entries).

Verify it's alive

curl https://modelishub.com/v1/chat/completions \
  -H "Authorization: Bearer $MODELIS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"modelis-auto","messages":[{"role":"user","content":"Reply with exactly: ok"}]}'

---

Model picker

Tier selection uses these colon names. modelis-auto is the standard tier.

Write this	Within-limit price	Included tokens
modelis-auto or modelis-auto:standard	$0.04 per call	8k input / 2k output
modelis-auto:basic	$0.008 per call	4k input / 1k output
modelis-auto:premium	$0.10 per call	4k input / 1k output

A call that stays inside that tier's input and output limits pays the per-call price. Usage above those limits is billed per token.

Named models are requested with their upstream model id and billed at the upstream price + 15%. The 15% is applied at top-up.

Streaming and tool-calling both work — the two things agentic tools actually need.

Why not just use the vendor's own key?

• One key, ~95 models — no juggling separate OpenAI/Anthropic/Google keys as you switch tasks.
• Per-call pricing on modelis-auto — each request within the tier's token allowance costs the same, so retry loops are easy to budget. Usage above that allowance is billed per token. Named models are upstream price + 15% (applied at top-up).
• Zero migration — OpenAI-compatible wire protocol, so it's a base URL swap to https://modelishub.com/v1, not a rewrite.

Not affiliated with Roo Code / Zoo Code / Cline / Continue — this is an independent config repo for people migrating providers.

License

MIT — copy anything here into your own dotfiles.