How do I connect SillyTavern to Claude or GPT with ApiTopMix?

In SillyTavern, open the API Connections panel, choose API = Chat Completion, set the source to Custom (OpenAI-compatible), enter https://apitopmix.com/v1 as the endpoint, paste your ApiTopMix API key, type a model ID such as claude-opus-4-6 or claude-sonnet-4-6, then click Connect. Because ApiTopMix exposes an OpenAI-compatible endpoint, no reverse-proxy hacks are needed.

Which Claude model is best for SillyTavern roleplay?

claude-opus-4-6 gives the richest, most coherent long-form roleplay; claude-sonnet-4-6 is the best balance of quality and cost for everyday chats; claude-haiku-4-5-20251001 is the cheapest for high-volume or simple scenes. You can switch models any time by changing the model ID in SillyTavern.

Is ApiTopMix cheaper than the official Anthropic API?

Yes. ApiTopMix prices Claude at roughly 60-70% of Anthropic's official rates (about 30-40% off). For example Claude Sonnet 4.6 is $2.20 input / $11 output per 1M tokens versus Anthropic's $3 / $15, and Opus 4.6 is $3 / $15 versus $5 / $25. See the live pricing page for current rates.

ApiTopMix does not store request or response bodies. It keeps only de-identified call metadata (timestamp, model, token count, status code) for billing and troubleshooting. Usage is subject to each upstream model provider's policies and ApiTopMix's terms. See the privacy policy for details.

Guide

How to Use Claude & GPT in SillyTavern (OpenAI-Compatible Setup)

Published: 2026-06-30 · ~6 min read · By ApiTopMix

If you run SillyTavern, you already know the pain: the official Claude and OpenAI APIs are expensive, rate-limited, and you end up juggling separate accounts and billing for each provider. This guide shows you how to drive Claude (Opus, Sonnet, Haiku) and GPT inside SillyTavern through a single OpenAI-compatible API key — at discounted, pay-per-use pricing, with no monthly fee.

It takes about 3 minutes and needs no reverse-proxy hacks, because ApiTopMix exposes a native OpenAI-compatible endpoint that SillyTavern speaks out of the box.

Why use ApiTopMix with SillyTavern

One key, many models. Switch between Claude Opus/Sonnet/Haiku, GPT, Gemini and DeepSeek by changing one field — no new account per provider.
Cheaper Claude. Claude is priced at roughly 60–70% of Anthropic's official rates (about 30–40% off). Heavy roleplay sessions burn a lot of tokens, so this adds up fast.
Pay-per-use, no subscription. You pay only for the tokens you actually use. Minimum top-up is $5, no monthly fee, no token expiry.
OpenAI-compatible. SillyTavern's built-in "Custom (OpenAI-compatible)" source connects directly — no proxy server to host yourself.

Step-by-step: connect SillyTavern to ApiTopMix

1. Get your API key

Create an account at apitopmix.com/register.
Top up (minimum $5).
Go to the console → Tokens / API Keys → create a key and copy the sk-... string.

2. Configure SillyTavern

Open SillyTavern, click the plug icon (API Connections) in the top bar, and set:

API: Chat Completion
Chat Completion Source: Custom (OpenAI-compatible)
Custom Endpoint (Base URL): https://apitopmix.com/v1
Custom API Key: your sk-... key from step 1
Enter a Model ID: e.g. claude-opus-4-6 or claude-sonnet-4-6

Click Connect. When the indicator turns green, send a test message. That's it — you're now talking to Claude through ApiTopMix.

Note: Because the endpoint is OpenAI-compatible, you do not need SillyTavern's Claude reverse-proxy fields. Just use the Custom (OpenAI-compatible) source above. Your existing presets, prompts and context settings stay exactly the same.

3. Pick a model

Change the model any time by editing the model ID field — no reconnect needed.

Model ID	Best for
`claude-opus-4-6`	Richest, most coherent long-form roleplay; best character consistency
`claude-sonnet-4-6`	Best quality-to-cost balance for everyday chats
`claude-haiku-4-5-20251001`	Cheapest & fastest for high-volume or simple scenes
`claude-opus-4-6-thinking`	Extended reasoning for complex, multi-thread scenarios

The full model list and live prices are on the pricing page.

Pricing example

Roleplay is token-heavy, so the per-token rate matters more than for most apps. ApiTopMix prices Claude below Anthropic's official rates:

Model	ApiTopMix (in / out, 1M tokens)	Anthropic official	Save
Claude Sonnet 4.6	$2.20 / $11.00	$3.00 / $15.00	~27%
Claude Opus 4.6	$3.00 / $15.00	$5.00 / $25.00	40%
Claude Haiku 4.5	$0.60 / $3.00	$1.00 / $5.00	40%

Billing: pay-per-use, billed on actual tokens consumed. No subscription, no minimum monthly spend, no token expiry. Minimum first top-up is $5. Always check the pricing page for current rates.

FAQ

Does streaming work?

Yes. ApiTopMix implements the standard OpenAI streaming (SSE) protocol, so SillyTavern's streaming toggle works normally.

Can I use group chats and presets?

Yes. Anything that works with an OpenAI-compatible source in SillyTavern — group chats, character cards, presets, author's notes, world info — works unchanged. ApiTopMix only changes where the request is sent, not how SillyTavern builds it.

What about latency?

Requests route through the upstream provider's official channel, so time-to-first-token is close to connecting directly (typically ~300–800ms depending on your region and network).

Are my chats stored?

ApiTopMix does not store request or response bodies. Only de-identified call metadata (timestamp, model, token count, status code) is kept for billing and troubleshooting. Usage is subject to each upstream model provider's policies and ApiTopMix's terms. See the privacy policy.

Can I also use GPT, Gemini or DeepSeek the same way?

Yes — same endpoint and key, just change the model ID (for example a GPT or Gemini model name). One account covers all supported providers.