Freeseek FREESEEK
SIGN IN GET FREE API KEY ↗
INTEGRATION TUTORIAL STEP-BY-STEP

CONNECTING JANITORAI TO
DEEPSEEK FLASH

Step-by-step instructions to connect your JanitorAI characters directly to DeepSeek Flash (enter freeseek-flash in JanitorAI) in under 2 minutes.

01

ARCHITECTURE & DEFINITIONS

OVERVIEW

WHAT IS A "MODEL"?

In JanitorAI, a Model is the AI engine that generates character replies, emotion, and story progression. Connecting to DeepSeek Flash delivers instantaneous streaming replies, sharper character dialogue, and rich descriptive vocabulary straight from DeepSeek.

WHAT IS A "REVERSE PROXY"?

A Reverse Proxy is a high-speed secure gateway. It accepts OpenAI-formatted requests from JanitorAI, attaches necessary auth tokens, and relays them straight from DeepSeek.

// TARGET: DEEPSEEK UPSTREAM ZERO BROKER HOPS LATENCY: LOW
02

2-STAGE SETUP SEQUENCE

CONFIGURATION
STAGE 1: GENERATE YOUR FREE PERSONAL API KEY
  1. Navigate to /register and create a free account (10 seconds).
  2. In your User Dashboard, click COPY KEY (prefixed with fsk-...).
STAGE 2: CONFIGURE JANITORAI REVERSE PROXY
  1. Open JanitorAI chat → API Settings → Custom OpenAI Proxy.
  2. Configure endpoint fields:
    PROXY URL
    MODEL NAME
    API KEY
  3. Click Save Settings & Refresh the page tab.
// MODEL == freeseek-flash PROTOCOL: OPENAI V1
03

RECOMMENDED CHAT & MEMORY

OPTIMIZATIONS

MODEL SELECTION: DEEPSEEK FLASH

Freeseek is dedicated exclusively to DeepSeek Flash (enter freeseek-flash as model name), powered straight from DeepSeek.

RECOMMENDED CONTEXT SIZE: 16K

In JanitorAI Chat Settings, configure Context Size to 16,384 (16k). This gives characters deep long-term memory of previous narrative beats without compromising streaming speed.

GENERATION LIMIT: 500-800

Set Max Generation Tokens to 600 for optimal pacing, paragraph formatting, and natural roleplay dialogue cadence.

// RECOMMENDED CONTEXT: 16,384 MAX GEN: 600 TOKENS
04

DIAGNOSTICS & ERROR CODES

DIAGNOSTICS
  • HTTP 401 UNAUTHORIZED (INVALID API KEY):
    Ensure you have created an account at /register and copied your complete fsk-... key without trailing spaces.
  • HTTP 403 FORBIDDEN (ROLEPLAYING SERVICE ONLY):
    This service is exclusively configured for character roleplaying on JanitorAI. Non-roleplaying queries are restricted.
  • HTTP 400 BAD REQUEST (MODEL MISMATCH):
    Ensure your custom model name in JanitorAI is set to freeseek-flash.
  • NO RESPONSE FROM CHARACTER:
    Re-open JanitorAI API Settings, click Save Settings again, and perform a full page refresh in your browser.
// TELEMETRY: 200 OK / 401 UNAUTH / 403 FORBIDDEN / 400 BAD REQ LOGS MONITORED
CREATE FREE ACCOUNT & GET API KEY ↗