PORDL SillyTavern Setup

Connect SillyTavern, RisuAI, or any OpenAI-compatible chat app to PORDL in under 5 minutes. Stream DeepSeek V4 and GPT models with no content filters and no per-token markup.

Why PORDL for SillyTavern

If you're coming from Character.AI, Janitor AI, or CrushOn, here's what changes:

• No content filters — PORDL passes your requests to DeepSeek and GPT models without adding moderation layers
• No surprise bills — flat monthly pricing instead of unpredictable per-token charges
• No 5.5% markup — unlike OpenRouter, PORDL charges 0% on DeepSeek models
• Model choice — switch between DeepSeek V4 Flash, DeepSeek V4 Pro, and GPT models from the SillyTavern dropdown
• Streaming — see text generate in real time, essential for immersive roleplay

Cost Comparison

Setup Guide

Step 1: Create Your PORDL Account

Go to https://pordl.dev and click Start Free.

You'll get a free account with 100K tokens/month — enough for several test sessions to make sure everything works before committing to a paid plan.

Copy your API key (starts with pd_live_). You'll need it in step 3.

Step 2: Install SillyTavern (if you haven't)

If SillyTavern is already installed, skip to Step 3.

SillyTavern requires Node.js. Install it first:

Windows:

# Download Node.js from https://nodejs.org (LTS version)
# Then in a terminal:
git clone https://github.com/SillyTavern/SillyTavern.git
cd SillyTavern
npm install
node server.js

Mac/Linux:

git clone https://github.com/SillyTavern/SillyTavern.git
cd SillyTavern
npm install
node server.js

Open your browser to http://localhost:8000.

Step 3: Connect SillyTavern to PORDL

1. In SillyTavern, click the API Connection panel (plug icon, top bar)
2. Set API Type to Custom OpenAI
3. Set Custom Endpoint (Base URL) to:

   https://api.pordl.dev/v1
   

4. Paste your PORDL API key in the API Key field
5. Click Connect — you should see a green checkmark

Step 4: Pick Your Model

Click the Model dropdown. You should see:

• deepseek-v4-flash — Best value. Great prose, fast responses, $0.14/M tokens. Start here.
• deepseek-v4-pro — Premium quality. Richer character voices, deeper reasoning. $0.435/M tokens.
• gpt-4o-mini — Fast and cheap OpenAI option.
• gpt-4o — Full GPT quality when you want it.
• auto — Let PORDL pick the best model for each message automatically.

Recommendation: Start with deepseek-v4-flash. It handles roleplay, character chat, and creative writing beautifully at the lowest cost. Switch to deepseek-v4-pro for your most important scenes.

Step 5: Configure for Best Results

Streaming: Make sure streaming is ON (it should be by default). This lets you see text appear as the model writes — much better for immersive roleplay.

Context Size: Set to match your model:

• DeepSeek V4 Flash/Pro: up to 1,000,000 (but 8K-32K is practical for most sessions)
• GPT-4o: up to 128,000

Temperature: 0.85–1.0 for creative roleplay. Below 0.7 gets repetitive. Above 1.2 gets incoherent.

Max Response Length: 400–600 tokens is a good starting point. Keeps pacing tight without cutting off mid-thought.

Step 6: Send a Test Message

Create or import a character card and send a message. You should see:

• Text streaming in real-time
• The character responding in voice
• No content filter interruptions

If it works, you're done. Start creating.

RisuAI Setup

RisuAI follows a similar process:

1. Open RisuAI settings
2. Go to Provider Settings
3. Select Custom OpenAI as the provider
4. Set the API URL to https://api.pordl.dev/v1
5. Enter your PORDL API key
6. Select deepseek-v4-flash as your model
7. Enable streaming

Other OpenAI-Compatible Apps

Any app that supports a custom OpenAI endpoint works with PORDL. The universal setup is:

• Base URL: https://api.pordl.dev/v1
• API Key: Your pd_live_ key from pordl.dev
• Model: deepseek-v4-flash (or any model from /v1/models)

This includes apps like Moltbook, Maid, Open WebUI, and any tool built on the OpenAI SDK.

Troubleshooting

"Could not connect" or "Invalid API key"

• Double-check your API key — it should start with pd_live_
• Make sure the base URL is exactly https://api.pordl.dev/v1 (with the /v1 at the end)
• Check that your account is active at https://pordl.dev

Responses are slow

• Switch to deepseek-v4-flash — it's the fastest model
• Reduce your context size if it's set very high
• Make sure streaming is enabled — without it, you wait for the entire response before seeing anything

Character "forgets" things mid-conversation

• Your context window may be too small — increase it in SillyTavern settings
• DeepSeek V4 supports up to 1M tokens of context, so you have room
• Use SillyTavern's summary/memory features to compress older conversation history

Hit the token limit

• Upgrade your plan at https://pordl.dev
• Creator ($4.99/mo) gives you 1M tokens — enough for 50-70 full roleplay sessions
• Creator Pro ($9.99/mo) gives you 5M tokens for heavy daily use

Plans

Links

• Sign up: https://pordl.dev
• SillyTavern: https://github.com/SillyTavern/SillyTavern
• RisuAI: https://risuai.net
• PORDL source: https://github.com/plagtech/pordl
• PORDL models: https://api.pordl.dev/v1/models