fastapi-guard is a security library for FastAPI that provides middleware to control IPs, log requests, detect penetration attempts and more. It integrates seamlessly with FastAPI to offer robust protection against various security threats.
🌐 Website - Check out the website!
🎮 Join our Discord Community - Connect with other developers!
📚 Documentation - Full technical documentation and deep dive into its inner workings.
🕹️ Live Playground - Testing playground for FastAPI Guard's features in action.
🤖 Monitoring Agent Integration - Monitor your FastAPI Guard instance with a monitoring agent.
- IP Whitelisting and Blacklisting: Control access based on IP addresses.
- User Agent Filtering: Block requests from specific user agents.
- Rate Limiting: Limit the number of requests from a single IP.
- Automatic IP Banning: Automatically ban IPs after a certain number of suspicious requests.
- Penetration Attempt Detection: Detect and log potential penetration attempts.
- Custom Logging: Log security events to a custom file.
- CORS Configuration: Configure CORS settings for your FastAPI application.
- Cloud Provider IP Blocking: Block requests from cloud provider IPs (AWS, GCP, Azure).
- IP Geolocation: Use a service like IPInfo.io API to determine the country of an IP address.
- Distributed State Management: (Optional) Redis integration for shared security state across instances
- Flexible Storage: Redis-enabled distributed storage or in-memory storage for single instance deployments
To install fastapi-guard, use pip:
pip install fastapi-guardTo use fastapi-guard, you need to configure the middleware in your FastAPI application. Here's a basic example:
from fastapi import FastAPI
from guard.middleware import SecurityMiddleware
from guard.models import SecurityConfig
from guard.handlers.ipinfo_handler import IPInfoManager
app = FastAPI()
# Define your security configuration
config = SecurityConfig(
ipinfo_token="your_ipinfo_token_here", # Optional: IPInfo token required for IP geolocation
ipinfo_db_path="custom/ipinfo.db", # Optional custom database path
whitelist=["192.168.1.1", "2001:db8::1"],
blacklist=["10.0.0.1", "2001:db8::2"],
blocked_countries=["AR", "IT"],
blocked_user_agents=["curl", "wget"],
auto_ban_threshold=5,
auto_ban_duration=86400,
custom_log_file="security.log",
rate_limit=100,
enforce_https=True,
enable_cors=True,
cors_allow_origins=["*"],
cors_allow_methods=["GET", "POST"],
cors_allow_headers=["*"],
cors_allow_credentials=True,
cors_expose_headers=["X-Custom-Header"],
cors_max_age=600,
block_cloud_providers={"AWS", "GCP", "Azure"},
)
# Add the security middleware
app.add_middleware(SecurityMiddleware, config=config)
@app.get("/")
async def read_root():
return {"message": "Hello, World!"}You can control access based on IP addresses using the whitelist and blacklist options in the SecurityConfig.
config = SecurityConfig(
whitelist=["192.168.1.1", "2001:db8::1"],
blacklist=["10.0.0.1", "2001:db8::2"],
)Block requests from specific user agents by adding patterns to the blocked_user_agents list.
config = SecurityConfig(
blocked_user_agents=["curl", "wget"],
)Limit the number of requests from a single IP using the rate_limit option.
config = SecurityConfig(
rate_limit=100, # Maximum 100 requests per minute
)Automatically ban IPs after a certain number of suspicious requests using the auto_ban_threshold and auto_ban_duration options.
config = SecurityConfig(
auto_ban_threshold=5, # Ban IP after 5 suspicious requests
auto_ban_duration=86400, # Ban duration in seconds (1 day)
)Enable penetration attempt detection using the enable_penetration_detection option.
config = SecurityConfig(
enable_penetration_detection=True, # True by default
)Optional: Enable passive mode to log suspicious activity without blocking requests.
config = SecurityConfig(
passive_mode=True, # False by default
)Detect and log potential penetration attempts using the detect_penetration_attempt function.
from fastapi import Request, Response, status
from fastapi.responses import JSONResponse
from guard.utils import detect_penetration_attempt
@app.post("/submit")
async def submit_data(request: Request):
is_suspicious, trigger_info = await detect_penetration_attempt(request)
if is_suspicious:
return JSONResponse(
status_code=status.HTTP_400_BAD_REQUEST,
content={"error": f"Suspicious activity detected: {trigger_info}"}
)
return {"message": "Data submitted successfully"}Log security events with console output (always enabled) and optional file logging:
config = SecurityConfig(
custom_log_file="security.log", # Optional: adds file logging
# custom_log_file=None, # Default: console output only
)Note: Console output is always enabled for visibility. File logging is only activated when custom_log_file is provided.
Configure CORS settings for your FastAPI application using the enable_cors and related options.
config = SecurityConfig(
enable_cors=True,
cors_allow_origins=["*"],
cors_allow_methods=["GET", "POST"],
cors_allow_headers=["*"],
cors_allow_credentials=True,
cors_expose_headers=["X-Custom-Header"],
cors_max_age=600,
)Block requests from cloud provider IPs (AWS, GCP, Azure) using the block_cloud_providers option.
config = SecurityConfig(
block_cloud_providers={"AWS", "GCP", "Azure"},
)If you want to use fastapi-guard's built-in country filtering features, you'll need to obtain an IPInfo token:
- Visit IPInfo's website to create a free account
- After signing up, you'll receive an API token
- The free tier includes:
- Up to 50,000 requests per month
- Access to IP to Country database
- Daily database updates
- IPv4 & IPv6 support
Note: This is only required if you use country filtering (blocked_countries, whitelist_countries). You can also provide your own handler that uses any other service.
FastAPI Guard provides powerful decorators that allow you to apply security controls to individual routes, giving you fine-grained control over your API endpoints.
from fastapi import FastAPI
from guard.middleware import SecurityMiddleware
from guard.models import SecurityConfig
from guard.decorators import SecurityDecorator
app = FastAPI()
config = SecurityConfig()
# Create decorator instance
guard_deco = SecurityDecorator(config)
# Apply decorators to specific routes
@app.get("/api/public")
def public_endpoint():
return {"data": "public"}
@app.get("/api/limited")
@guard_deco.rate_limit(requests=10, window=300) # 10 requests per 5 minutes
def limited_endpoint():
return {"data": "limited"}
@app.get("/api/restricted")
@guard_deco.require_ip(whitelist=["192.168.1.0/24"])
@guard_deco.block_countries(["CN", "RU"])
def restricted_endpoint():
return {"data": "restricted"}
# Add global middleware
app.add_middleware(SecurityMiddleware, config=config)
# Required: Set decorator handler on app state
app.state.guard_decorator = guard_decoAccess Control
@guard_deco.require_ip(whitelist=[], blacklist=[])- IP address filtering@guard_deco.block_countries(["CN", "RU"])- Block specific countries@guard_deco.allow_countries(["US", "CA"])- Allow only specific countries@guard_deco.block_clouds(["AWS", "GCP"])- Block cloud provider IPs
Rate Limiting
@guard_deco.rate_limit(requests=10, window=60)- Basic rate limiting@guard_deco.geo_rate_limit(limits={"US": 100, "default": 50})- Geographic rate limiting
Authentication & Headers
@guard_deco.require_https()- Force HTTPS@guard_deco.require_auth(type="bearer")- Require authentication@guard_deco.api_key_auth(header_name="X-API-Key")- API key authentication@guard_deco.require_headers({"X-Custom": "required"})- Require specific headers
Content Filtering
@guard_deco.block_user_agents(["curl", "wget"])- Block user agent patterns@guard_deco.content_type_filter(["application/json"])- Filter content types@guard_deco.max_request_size(1048576)- Limit request size (1MB)@guard_deco.require_referrer(["myapp.com"])- Require specific referrers
Behavioral Analysis
@guard_deco.usage_monitor(max_calls=50, window=3600, action="ban")- Monitor endpoint usage@guard_deco.return_monitor("rare_item", max_occurrences=3, window=86400, action="alert")- Monitor return patterns@guard_deco.suspicious_frequency(max_frequency=0.1, window=300, action="log")- Detect suspicious frequency
Advanced Controls
@guard_deco.time_window("09:00", "17:00", "UTC")- Time-based access control@guard_deco.honeypot_detection(trap_fields=["hidden_field"])- Detect bots using honeypot fields@guard_deco.bypass(checks=["rate_limit"])- Bypass specific security checks
Combine multiple decorators for comprehensive protection:
@app.post("/api/admin/sensitive")
@guard_deco.require_https() # Security requirement
@guard_deco.require_auth(type="bearer") # Authentication
@guard_deco.require_ip(whitelist=["10.0.0.0/8"]) # Access control
@guard_deco.rate_limit(requests=5, window=3600) # Rate limiting
@guard_deco.suspicious_detection(enabled=True) # Monitoring
def admin_endpoint():
return {"status": "admin action"}
@app.get("/api/rewards")
@guard_deco.usage_monitor(max_calls=50, window=3600, action="ban")
@guard_deco.return_monitor("rare_item", max_occurrences=3, window=86400, action="ban")
@guard_deco.block_countries(["CN", "RU", "KP"])
def rewards_endpoint():
# This endpoint is protected against:
# - Excessive usage (>50 calls/hour results in ban)
# - Suspicious return patterns (>3 rare items/day results in ban)
# - Geographic restrictions
return {"reward": "rare_item", "value": 1000}Security settings are applied in the following priority order:
- Decorator Settings (highest priority)
- Global Middleware Settings
- Default Settings (lowest priority)
This allows routes to override global settings while maintaining sensible defaults.
Configure trusted proxies to securely handle X-Forwarded-For headers:
config = SecurityConfig(
trusted_proxies=["10.0.0.1", "192.168.1.0/24"], # List of trusted proxy IPs or CIDR ranges
trusted_proxy_depth=1, # How many proxies to expect in chain
trust_x_forwarded_proto=True, # Whether to trust X-Forwarded-Proto for HTTPS detection (default: True)
)When trusted_proxies is configured, FastAPI Guard will:
- Only trust X-Forwarded-For headers from these IPs
- Extract the appropriate client IP based on proxy depth
- Prevent IP spoofing attacks through header manipulation
The library implements a handler that uses IPInfo's IP to Country database, which provides:
- Full accuracy IP to country mapping
- Daily updates
- Support for both IPv4 and IPv6
- Country and continent information
- ASN details
To use the geolocation feature with this handler:
from guard.protocols.geoip_handler import GeoIPHandler
config = SecurityConfig(
geo_ip_handler=GeoIPHandler,
blocked_countries=["AR", "IT"], # Block specific countries using ISO 3166-1 alpha-2 codes
whitelist_countries=["US", "CA"] # Optional: Only allow specific countries
)The database is automatically downloaded and cached locally when the middleware starts, if required, and it's updated daily to ensure accuracy.
You can also use a service other than IPInfo, as long as you implement the same protocol:
# implement the required methods of guard.protocols.geoip_handler.GeoIPHandler protocol
class GeoIPHandler:
"""
Your custom class.
"""
@property
def is_initialized(self) -> bool:
# your implementation
...
async def initialize(self) -> None:
# your implementation
...
async def initialize_redis(self, redis_handler: "RedisManager") -> None:
# your implementation
...
def get_country(self, ip: str) -> str | None:
# your implementation
...
config = SecurityConfig(
geo_ip_handler=GeoIPHandler,
blocked_countries=["AR", "IT"], # Block specific countries using ISO 3166-1 alpha-2 codes
whitelist_countries=["US", "CA"] # Optional: Only allow specific countries
)You can define a custom function to perform additional checks on the request using the custom_request_check option.
from fastapi import Request, Response
async def custom_check(request: Request) -> Optional[Response]:
if "X-Custom-Header" not in request.headers:
return Response("Missing custom header", status_code=400)
return None
config = SecurityConfig(
custom_request_check=custom_check,
)You can define a custom function to modify the response before it's sent using the custom_response_modifier option.
from fastapi import Response
from fastapi.responses import JSONResponse
async def custom_modifier(response: Response) -> Response:
# Add custom headers
response.headers["X-Custom-Header"] = "CustomValue"
# Convert text responses to FastAPI-style JSON responses
if response.status_code >= 400 and not isinstance(response, JSONResponse):
try:
content = response.body.decode()
return JSONResponse(
status_code=response.status_code,
content={"detail": content}
)
except:
pass
return response
config = SecurityConfig(
custom_response_modifier=custom_modifier,
)The example above shows how to:
- Add custom headers to all responses
- Convert plain text error responses to JSON format with a "detail" field, matching FastAPI's HTTPException format
Enable distributed state management across multiple instances:
config = SecurityConfig(
enable_redis=True,
redis_url="redis://prod-redis:6379/1",
redis_prefix="myapp:security:",
)The Redis integration provides:
- Atomic increment operations for rate limiting
- Distributed IP ban tracking
- Cloud provider IP range caching
- Pattern storage for penetration detection
The SecurityConfig class defines the structure for security configuration, including IP whitelists and blacklists, blocked countries, blocked user agents, rate limiting, automatic IP banning, HTTPS enforcement, custom hooks, CORS settings, and blocking of cloud provider IPs.
- geo_ip_handler:
GeoIPHandler- Protocol that allows for IP geolocation functionality. - enable_redis:
bool- Enable Redis for distributed state (default: True). When disabled, uses in-memory storage. - redis_url:
str | None- Redis connection URL (default: "redis://localhost:6379"). - redis_prefix:
str- Prefix for Redis keys (default: "fastapi_guard:"). - trusted_proxies:
list[str] | None- List of trusted proxy IPs or CIDR ranges. - trusted_proxy_depth:
int- How many proxies to expect in chain. - trust_x_forwarded_proto:
bool- Whether to trust X-Forwarded-Proto for HTTPS detection. - whitelist:
list[str] | None- A list of IP addresses or ranges that are always allowed. If set to None, no whitelist is applied. - blacklist:
list[str]- A list of IP addresses or ranges that are always blocked. - blocked_countries:
list[str]- A list of country codes whose IP addresses should be blocked. - blocked_user_agents:
list[str]- A list of user agent strings or patterns that should be blocked. - auto_ban_threshold:
int- The threshold for auto-banning an IP address after a certain number of requests. - auto_ban_duration:
int- The duration in seconds for which an IP address should be banned after reaching the auto-ban threshold. - custom_log_file:
str | None- Optional path to a log file. When provided, enables file logging in addition to console output (which is always enabled). Default:None(console only). - custom_error_responses:
dict[int, str]- A dictionary of custom error responses for specific HTTP status codes. - rate_limit:
int- The maximum number of requests allowed per minute from a single IP. - enforce_https:
bool- Whether to enforce HTTPS connections. If True, all HTTP requests will be redirected to HTTPS. - custom_request_check:
Callable[[Request], Awaitable[Response | None]] | None- A custom function to perform additional checks on the request. If it returns a Response, that response will be sent instead of continuing the middleware chain. - custom_response_modifier:
Callable[[Response], Awaitable[Response]] | None- A custom function to modify the response before it's sent. - enable_cors:
bool- Whether to enable CORS. - cors_allow_origins:
list[str]- A list of origins that are allowed to access the API. - cors_allow_methods:
list[str]- A list of methods that are allowed to access the API. - cors_allow_headers:
list[str]- A list of headers that are allowed in CORS requests. - cors_allow_credentials:
bool- Whether to allow credentials in CORS requests. - cors_expose_headers:
list[str]- A list of headers that are exposed in CORS responses. - cors_max_age:
int- The maximum age in seconds that the results of a preflight request can be cached. - block_cloud_providers:
set[str]- Case-sensitive cloud provider names to block. Valid values: 'AWS', 'GCP', 'Azure'. Invalid entries are silently ignored. - ipinfo_token:
str(DEPRECATED) - The IPInfo API token required for IP geolocation functionality. - ipinfo_db_path:
str(DEPRECATED) - Custom path for IPInfo database storage (default: ./data/ipinfo/country_asn.mmdb)
Contributions are welcome! Please open an issue or submit a pull request on GitHub.
This project is licensed under the MIT License. See the LICENSE file for details.
Renzo Franceschini - [email protected]
