Server Configuration

Host

Port

Protocol

Public Domain (for Let's Encrypt) If specified, Let's Encrypt will be used to generate valid SSL certificates. Leave empty to use self-signed certificates.

SSL Certificate Path Path to SSL certificate file. Leave empty to use default (~/.aisbf/cert.pem). Auto-generated using Let's Encrypt if public domain is set, otherwise self-signed.

SSL Key Path Path to SSL private key file. Leave empty to use default (~/.aisbf/key.pem). Auto-generated with certificate.

Authentication

Enable API Authentication

Auth Tokens (one per line)

MCP Server

Enable MCP Server Model Context Protocol server for remote agent configuration

Autoselect Tokens (one per line) Tokens that give access to autoselection and autorotation settings + standard APIs

Full Config Tokens (one per line) Tokens that give access to full system configuration + standard APIs

Internal Models

Condensation Model ID

Used when condensation model is set to "internal"

Condensation Max Tokens Max tokens the condensation model can generate per summary (default 1000).

Autoselect Model ID

Used when autoselect selection_model is set to "internal"

Autoselect Context Limit (tokens) Max tokens of conversation context sent to the internal autoselect model (default 8000). Ignored when selection_model is a rotation or provider/model.

NSFW Classifier Model ID

Model used for NSFW content detection

Privacy Classifier Model ID

Model used for privacy-sensitive information detection

Semantic Vectorization Model ID

Model used for semantic embedding and vectorization

Removes all locally cached model files and unloads them from memory. Models will be re-downloaded from HuggingFace on next use.

Database Configuration

Database Type Choose the database backend (SQLite is recommended for most users)

SQLite Database Path Path to SQLite database file (supports ~ expansion)

MySQL Host MySQL server hostname or IP address

MySQL Port MySQL server port (default: 3306)

MySQL Username MySQL database username

MySQL Password MySQL database password

MySQL Database Name MySQL database name

Cache Configuration

Cache Type Choose the cache backend for model embeddings and other cached data (SQLite recommended for most users)

SQLite Cache Path Path to SQLite cache database file (supports ~ expansion)

MySQL Host MySQL server hostname or IP address

MySQL Port MySQL server port (default: 3306)

MySQL Username MySQL database username

MySQL Password MySQL database password

MySQL Database Name MySQL database name for cache

Redis Host Redis server hostname or IP address

Redis Port Redis server port (default: 6379)

Redis Database Redis database number (default: 0)

Redis Password Redis password (optional)

Redis Key Prefix Prefix for Redis keys (default: aisbf:)

Response Cache Configuration

Enable Response Cache Cache API responses to improve performance and reduce costs

Response Cache Backend Choose the backend for response caching

Cache TTL (seconds) Time to live for cached responses (default: 600 seconds)

Max Memory Cache Size Maximum number of items in memory cache (default: 1000)

Redis Host

Redis Port

Redis Database

Redis Password

Redis Key Prefix

SQLite Path

MySQL Host

MySQL Port

MySQL Username

MySQL Password

MySQL Database

Response Cache

Status: Loading...

Content Classification

Enable NSFW Classification Enable automatic NSFW content detection for model selection

Enable Privacy Classification Enable automatic privacy-sensitive content detection for model selection

Enable Semantic Classification Enable semantic content classification using the configured model

TOR Hidden Service

Status: Loading...

Onion Address:

Enable TOR Hidden Service Expose AISBF over TOR network as a hidden service

⚙️ Required torrc Configuration

Before enabling Tor, you must configure your torrc file (usually /etc/tor/torrc):

Step 1: Generate Password Hash

tor --hash-password "your_secure_password"

Step 2: Add to torrc

For temporary hidden service (new address each restart):

ControlPort 9051
HashedControlPassword 16:YOUR_HASHED_PASSWORD_HERE

⚠️ Leave "Hidden Service Directory" field blank below for ephemeral service

For permanent hidden service (same address across restarts):

ControlPort 9051
HashedControlPassword 16:YOUR_HASHED_PASSWORD_HERE

HiddenServiceDir /home/yourusername/.aisbf/tor_hidden_service
HiddenServicePort 80 127.0.0.1:{{ config.server.port if config.server else 17765 }}

⚠️ Replace /home/yourusername with your actual home directory path (torrc does NOT expand ~)

Step 3: Restart Tor

sudo systemctl restart tor  # Linux
brew services restart tor   # macOS

TOR Control Host TOR control port host (default: 127.0.0.1)

TOR Control Port TOR control port (default: 9051) - must match ControlPort in torrc

TOR Control Password The password you used with tor --hash-password (NOT the hash itself)

Hidden Service Directory

Ephemeral: Leave blank for temporary service (new address each restart)
Persistent: Enter full absolute path (e.g., /home/username/.aisbf/tor_hidden_service) - must match HiddenServiceDir in torrc
⚠️ Note: torrc does NOT expand ~ - use full paths like /home/username instead

Hidden Service Port Port exposed on the onion address (default: 80) - must match HiddenServicePort in torrc for persistent services

SOCKS Proxy Host TOR SOCKS proxy host (default: 127.0.0.1)

SOCKS Proxy Port TOR SOCKS proxy port (default: 9050)

User Signup

Enable User Signup Allow users to create accounts via signup form

Require Email Verification Users must verify their email before they can login

Verification Token Expiry (hours) How long verification links remain valid (1-168 hours)

OAuth2 Authentication

Google OAuth2

Enable Google OAuth2 Authentication

Google Client ID

Google Client Secret

GitHub OAuth2

Enable GitHub OAuth2 Authentication

GitHub Client ID

GitHub Client Secret

SMTP Email Configuration

Enable SMTP Email Enable email sending functionality for password resets, account verification, and notifications

SMTP Host SMTP server hostname (e.g., smtp.gmail.com)

SMTP Port SMTP server port (587 for TLS, 465 for SSL, 25 for no encryption)

SMTP Username SMTP authentication username (leave blank if not required)

SMTP Password SMTP authentication password

Use TLS Use STARTTLS for secure connection (recommended for port 587)

Use SSL Use SSL/TLS from the start (for port 465)

From Email Address Email address to send from

From Name Display name for sent emails

Request Batching

Enable Request Batching Batch multiple requests together for improved efficiency

Batching Window (milliseconds) Time window to collect requests for batching (default: 100ms)

Max Batch Size Maximum number of requests per batch (default: 8)

Provider-Specific Batching Settings

Enable OpenAI Batching

OpenAI Max Batch Size

Enable Anthropic Batching

Anthropic Max Batch Size

Client Rate Limiting

Protect the server against request flooding. Limits are applied per authenticated token (API/MCP) or per IP address (all other routes). Set a value to 0 to disable that specific threshold.

Enable Client Rate Limiting Rate-limit inbound clients before processing auth or routing

API & MCP Calls

Requests per Minute 0 = unlimited

Requests per Hour 0 = unlimited

General (Dashboard & Other)

Requests per Minute 0 = unlimited

Requests per Hour 0 = unlimited

Adaptive Rate Limiting

Outbound adaptive rate limiting — automatically adjusts how fast requests are forwarded to upstream providers based on their responses.

Enable Adaptive Rate Limiting Automatically adjust rate limits based on provider responses

Initial Rate Limit (requests/second) Starting rate limit (0 = unlimited)

Learning Rate How quickly to adapt to rate limit changes (0-1, default: 0.1)

Headroom Percent Safety margin below detected limit (default: 10%)

Recovery Rate Rate at which limits recover after errors (default: 0.05)

Max Rate Limit (requests/second) Maximum allowed rate limit (default: 60)

Min Rate Limit (requests/second) Minimum allowed rate limit (default: 0.1)

Backoff Base Exponential backoff multiplier (default: 2)

Jitter Factor Random jitter to prevent thundering herd (0-1, default: 0.25)

History Window (seconds) Time window for tracking rate limit history (default: 3600)

Consecutive Successes for Recovery Number of successful requests before increasing rate limit (default: 10)

Admin Account & Notifications

Admin Username

New Password

Confirm New Password

Admin Email Address Email address to receive admin notifications. Requires SMTP to be configured and enabled.

Admin Email Notifications

Select which events trigger an email to the admin. SMTP must be enabled and an admin email must be set above. {% set notif = config.dashboard.notifications if config.dashboard and config.dashboard.notifications else {} %}

New user self-registered Payment received Subscription upgrade Subscription downgrade Subscription expired / cancelled Subscription renewed Wallet top-up User deleted their account

Server Settings

Response Cache Configuration

Response Cache

⚙️ Required torrc Configuration

GitHub OAuth2

Provider-Specific Batching Settings