Skip to content
Arcjet Docs

Arcjet Bun SDK reference

npm badge

This is the reference guide for the Arcjet Bun SDK, available on GitHub and licensed under the Apache 2.0 license.

What is Arcjet? Arcjet helps developers protect their apps in just a few lines of code. Bot detection. Rate limiting. Email validation. Attack protection. Data redaction. A developer-first approach to security.

Installation

In your project root, run the following command to install the SDK:

Terminal window
bun add @arcjet/bun

Requirements

  • Bun 1.1.27 or later

Quick start

Check out the quick start guide.

Configuration

Create a new Arcjet object with your API key and rules.

The required fields are:

The optional fields are:

  • characteristics (string[]) - A list of characteristics to be used to uniquely identify clients.
  • proxies (string[]) - A list of one or more trusted proxies. These addresses will be excluded when Arcjet is determining the client IP address. This is useful if you are behind a load balancer or proxy that sets the client IP address in a header. See Load balancers & proxies below for an example.
import arcjet, { shield } from "@arcjet/bun";
import { env } from "bun";


export const aj = arcjet({
  key: env.ARCJET_KEY!, // Get your site key from https://app.arcjet.com
  rules: [
    // Protect against common attacks with Arcjet Shield
    shield({
      mode: "LIVE", // will block requests. Use "DRY_RUN" to log only
    }),
  ],
});

Single instance

We recommend creating a single instance of the Arcjet object and reusing it throughout your application. This is because the SDK caches decisions and configuration to improve performance.

The pattern we use is to create a utility file that exports the Arcjet object and then import it where you need it.

Rule modes

Each rule can be configured in either LIVE or DRY_RUN mode. When in DRY_RUN mode, each rule will return its decision, but the end conclusion will always be ALLOW.

This allows you to run Arcjet in passive / demo mode to test rules before enabling them. Arcjet will log what it would have done.

import arcjet, { fixedWindow } from "@arcjet/bun";
import { env } from "bun";


const aj = arcjet({
  key: env.ARCJET_KEY!,
  characteristics: ["ip.src"],
  rules: [
    fixedWindow(
      // This rule is live
      {
        mode: "LIVE",
        window: "1h",
        max: 60,
      },
      // This rule is in dry run mode, so will log but not block
      {
        mode: "DRY_RUN",
        characteristics: ['http.request.headers["x-api-key"]'],
        window: "1h",
        // max could also be a dynamic value applied after looking up a limit
        // elsewhere e.g. in a database for the authenticated user
        max: 600,
      },
    ),
  ],
});

Multiple rules

You can combine rules to create a more complex protection strategy. For example, you can combine rate limiting and bot protection rules to protect your API from automated clients.

index.ts
import arcjet, { detectBot, tokenBucket } from "@arcjet/bun";
import { env } from "bun";


// Create an Arcjet instance with multiple rules
const aj = arcjet({
  key: env.ARCJET_KEY!, // Get your site key from https://app.arcjet.com
  rules: [
    tokenBucket({
      mode: "LIVE", // will block requests. Use "DRY_RUN" to log only
      refillRate: 5, // refill 5 tokens per interval
      interval: 10, // refill every 10 seconds
      capacity: 10, // bucket maximum capacity of 10 tokens
    }),
    detectBot({
      mode: "LIVE",
      allow: [], // "allow none" will block all detected bots
    }),
  ],
});

Environment variables

  • ARCJET_BASE_URL - Will override the decision API which the SDK communicates with. This defaults to https://decide.arcjet.com and should only be changed if directed by Arcjet support.
  • ARCJET_LOG_LEVEL - The log level to use, either debug, info, warn, or error. Defaults to warn. If a rule is in dry run mode, a warning will be output with the decision that would have been applied.
  • ARCJET_ENV - Set to development to force Arcjet into development mode. This will allow private/internal addresses so that the SDKs work correctly locally. You usually do not need to set this because it uses NODE_ENV when set. See Troubleshooting for when this may be needed.

Custom logging

The SDK uses a lightweight logger which mirrors the Pino structured logger interface. You can use this to customize the logging output.

First, install the required packages:

Terminal window
bun install pino pino-pretty

Then, create a custom logger that will log to JSON in production and pretty print in development:

index.ts
import arcjet, { shield } from "@arcjet/bun";
import { env } from "bun";
import pino, { type Logger } from "pino";


const logger: Logger =
  env.ARCJET_ENV !== "development"
    ? // JSON in production, default to warn
      pino({ level: process.env.ARCJET_LOG_LEVEL || "warn" })
    : // Pretty print in development, default to debug
      pino({
        transport: {
          target: "pino-pretty",
          options: {
            colorize: true,
          },
        },
        level: process.env.ARCJET_LOG_LEVEL || "debug",
      });


const aj = arcjet({
  key: env.ARCJET_KEY!,
  rules: [
    // Protect against common attacks with Arcjet Shield
    shield({
      mode: "LIVE", // will block requests. Use "DRY_RUN" to log only
    }),
  ],
  // Use the custom logger
  log: logger,
});


export default {
  port: 3000,
  fetch: aj.handler(async (req) => {
    const decision = await aj.protect(req);


    if (decision.isDenied()) {
      return new Response("Access Denied", { status: 403 });
    }


    return new Response("Hello world");
  }),
};

Load balancers & proxies

If your application is behind a load balancer, Arcjet will only see the IP address of the load balancer and not the real client IP address.

To fix this, most load balancers will set the X-Forwarded-For header with the real client IP address plus a list of proxies that the request has passed through.

The problem with is that the X-Forwarded-For header can be spoofed by the client, so you should only trust it if you are sure that the load balancer is setting it correctly. See the MDN docs for more details.

You can configure Arcjet to trust IP addresses in the X-Forwarded-For header by setting the proxies field in the configuration. This should be a list of the IP addresses of your load balancers to be removed, so that the last IP address in the list is the real client IP address.

Example

For example, if the load balancer is at 100.100.100.100 and the client IP address is 192.168.1.1, the X-Forwarded-For header will be:

X-Forwarded-For: 192.168.1.1, 100.100.100.100

You should set the proxies field to ["100.100.100.100"] so Arcjet will use 192.168.1.1 as the client IP address.

import arcjet from "@arcjet/bun";
import { env } from "bun";


const aj = arcjet({
  key: env.ARCJET_KEY!,
  rules: [],
  proxies: ["100.100.100.100"],
});

Protect

Arcjet provides a single protect function that is used to execute your protection rules. This requires a request object which is the request argument as passed to the Bun fetch method. Rules you add to the SDK may require additional details, such as the validateEmail rule requiring an additional email prop.

This function returns a Promise that resolves to an ArcjetDecision object, which provides a high-level conclusion and detailed explanations of the decision made by Arcjet.

import arcjet, { shield } from "@arcjet/bun";
import { env } from "bun";


const aj = arcjet({
  key: env.ARCJET_KEY!, // Get your site key from https://app.arcjet.com
  rules: [
    // Protect against common attacks with Arcjet Shield
    shield({
      mode: "LIVE", // will block requests. Use "DRY_RUN" to log only
    }),
  ],
});


export default {
  port: 3000,
  fetch: aj.handler(async (req) => {
    const decision = await aj.protect(req);
    console.log("Arcjet decision", decision);


    if (decision.isDenied()) {
      return new Response("Forbidden", { status: 403 });
    }


    return new Response("Hello world");
  }),
};

aj.handler()

Arcjet uses client IP addresses for fingerprinting, but Bun doesn’t provide the IP address in the request object. By wrapping the fetch() handler in aj.handler(), the Arcjet SDK will be able to perform some preprocessing on the request to include the IP address.

You don’t need to use aj.handler() if you have another way of adding a proper IP address to the request object, or if there is an alternative way for the Arcjet SDK to detect the IP address, such as the Fly-Client-IP header on Fly.io.

Decision

The protect function function returns a Promise that resolves to an ArcjetDecision object. This contains the following properties:

  • id (string) - The unique ID for the request. This can be used to look up the request in the Arcjet dashboard. It is prefixed with req_ for decisions involving the Arcjet cloud API. For decisions taken locally, the prefix is lreq_.
  • conclusion ("ALLOW" | "DENY" | "CHALLENGE" | "ERROR") - The final conclusion based on evaluating each of the configured rules. If you wish to accept Arcjet’s recommended action based on the configured rules then you can use this property.
  • reason (ArcjetReason) - An object containing more detailed information about the conclusion.
  • results (ArcjetRuleResult[]) - An array of ArcjetRuleResult objects containing the results of each rule that was executed.
  • ttl (uint32) - The time-to-live for the decision in seconds. This is the time that the decision is valid for. After this time, the decision will be re-evaluated. The SDK automatically caches DENY decisions for the length of the TTL.
  • ip (ArcjetIpDetails) - An object containing Arcjet’s analysis of the client IP address. See IP analysis below for more information.

Conclusion

The ArcjetDecision object has the following methods that should be used to check the conclusion:

  • isAllowed() (bool) - The request should be allowed.
  • isDenied() (bool) - The request should be denied.
  • isErrored() (bool) - There was an unrecoverable error.

Reason

The reason property of the ArcjetDecision object contains an ArcjetReason object which provides more detailed information about the conclusion. This is the final decision reason and is based on the configured rules.

The ArcjetReason object has the following methods that can be used to check which rule caused the conclusion:

  • isBot() (bool) - Returns true if the bot protection rules have been applied and the request was considered to have been made by a bot.
  • isEmail() (bool) - Returns true if the email rules have been applied and the email address has a problem.
  • isRateLimit() (bool) - Returns true if the rate limit rules have been applied and the request has exceeded the rate limit.
  • isSensitiveInfo() (bool) - Returns true if sensitive info rules have been applied and sensitive info has been detected.
  • isShield() (bool) - Returns true if the shield rules have been applied and the request is suspicious based on analysis by Arcjet Shield.
  • isError() (bool) - Returns true if there was an error processing the request.

Results

The results property of the ArcjetDecision object contains an array of ArcjetRuleResult objects. There will be one for each configured rule so you can inspect the individual results:

  • id (string) - The ID of the rule result. Not yet implemented.
  • state (ArcjetRuleState) - Whether the rule was executed or not.
  • conclusion (ArcjetConclusion) - The conclusion of the rule. This will be one of the above conclusions: ALLOW, DENY, CHALLENGE, or ERROR.
  • reason (ArcjetReason) - An object containing more detailed information about the conclusion for this rule. Each rule type has its own reason object with different properties.

You can iterate through the results and check the conclusion for each rule.

for (const result of decision.results) {
  console.log("Rule Result", result);
}

This example will log the full result as well as each rate limit rule:

import arcjet, { detectBot, fixedWindow } from "@arcjet/bun";
import { env } from "bun";


const aj = arcjet({
  key: env.ARCJET_KEY!,
  // Tracking by ip.src is the default if not specified
  //characteristics: ["ip.src"],
  rules: [
    fixedWindow({
      mode: "LIVE",
      window: "1h",
      max: 60,
    }),
    detectBot({
      mode: "LIVE",
      allow: [], // "allow none" will block all detected bots
    }),
  ],
});


export default {
  port: 3000,
  fetch: aj.handler(async (req) => {
    const decision = await aj.protect(req);


    for (const result of decision.results) {
      console.log("Rule Result", result);


      if (result.reason.isRateLimit()) {
        console.log("Rate limit rule", result);
      }


      if (result.reason.isBot()) {
        console.log("Bot protection rule", result);
      }
    }


    if (decision.isDenied()) {
      return new Response("Forbidden", { status: 403 });
    }


    return new Response("Hello world");
  }),
};

Rule state

The state property of the ArcjetRuleResult object is an ArcjetRuleState. Each rule is evaluated individually and can be in one of the following states:

  • DRY_RUN - The rule was executed in dry run mode. This means that the rule was executed but the conclusion was not applied to the request. This is useful for testing rules before enabling them.
  • RUN - The rule was executed and the conclusion was applied to the request.
  • NOT_RUN - The rule was not executed. This can happen if another rule has already reached a conclusion that applies to the request. For example, if a rate limit rule is configured then these are evaluated before all other rules. If the client has reached the maximum number of requests then other rules will not be evaluated.
  • CACHED - The rule was not executed because the previous result was cached. Results are cached when the decision conclusion is DENY. Subsequent requests from the same client will not be evaluated against the rule until the cache expires.

Rule reason

The reason property of the ArcjetRuleResult object contains an ArcjetReason object which provides more detailed information about the conclusion for that configured rule.

Shield

The ArcjetReason object for shield rules has the following properties:

shieldTriggered: boolean;

See the shield documentation for more information about these properties.

Bot protection

The ArcjetReason object for bot protection rules has the following properties:

allowed: string[];
denied: string[];

Each of the allowed and denied arrays contains the identifiers of the bots allowed or denied from our full list of bots.

Rate limiting

The ArcjetReason object for rate limiting rules has the following properties:

max: number;
remaining: number;
window: number;
reset: number;

See the rate limiting documentation for more information about these properties.

Email validation & verification

The ArcjetReason object for email rules has the following properties:

emailTypes: ArcjetEmailType[];

An ArcjetEmailType is one of the following strings:

"DISPOSABLE" | "FREE" | "NO_MX_RECORDS" | "NO_GRAVATAR" | "INVALID";

See the email validation documentation for more information about these properties.

IP analysis

As of SDK version 1.0.0-alpha.11, the ArcjetDecision object contains an ip property. This includes additional data about the client IP address:

IP location

The following are available on the Free plan:

  • country (string | undefined): the country code the client IP address.
  • countryName (string | undefined): the country name of the client IP address.

The following are available on the Pro plan and above:

  • latitude (number | undefined): the latitude of the client IP address.
  • longitude (number | undefined): the longitude of the client IP address.
  • accuracyRadius (number | undefined): how accurate the location is in kilometers.
  • timezone (string | undefined): the timezone of the client IP address.
  • postalCode (string | undefined): the postal or zip code of the client IP address.
  • city (string | undefined): the city of the client IP address.
  • region (string | undefined): the region of the client IP address.
  • continent (string | undefined): the continent code of the client IP address.
  • continentName (string | undefined): the continent name of the client IP address.

The IP location fields may be undefined, but you can use various methods to check their availability. Using the methods will also refine the type to remove the need for null or undefined checks.

  • hasLatitude() (bool): returns whether the latitude and accuracyRadius fields are available.
  • hasLongitude() (bool): returns whether the longitude and accuracyRadius fields are available.
  • hasAccuracyRadius() (bool): returns whether the longitude, latitude, and accuracyRadius fields are available.
  • hasTimezone() (bool): returns whether the timezone field is available.
  • hasPostalCode() (bool): returns whether the postalCode field is available.
  • hasCity() (bool): returns whether the city field is available.
  • hasRegion() (bool): returns whether the region field is available.
  • hasCountry() (bool): returns whether the country and countryName fields are available.
  • hasContinent() (bool): returns whether the continent and continentName fields are available.
Location accuracy

IP geolocation can be notoriously inaccurate, especially for mobile devices, satellite internet providers, and even just normal users. Likewise with the specific fields like city and region, which can be very inaccurate. Country is usually accurate, but there are often cases where IP addresses are mis-located. These fields are provided for convenience e.g. suggesting a user location, but should not be relied upon by themselves.

IP AS

This is useful for identifying the network operator of the client IP address. This is useful for understanding whether the client is likely to be automated or not, or being stricter with requests from certain networks.

The IP AS fields may be undefined, but you can use the hasASN() method to check their availability. Using this method will also refine the type to remove the need for null-ish checks.

The following are available on the Pro plan and above:

  • hasASN() (bool): returns whether all of the ASN fields are available.
  • asn (string | undefined): the autonomous system (AS) number of the client IP address.
  • asnName (string | undefined): the name of the AS of the client IP address.
  • asnDomain (string | undefined): the domain of the AS of the client IP address.
  • asnType ('isp' | 'hosting' | 'business' | 'education'): the type of the AS of the client IP address. Real users are more likely to be on an ISP or business network rather than a hosting provider. Education networks often have a single or small number of IP addresses even though there are many users. A common mistake is to block a single IP because of too many requests when it is a university or company network using NAT (Network Address Translation) to give many users the same IP.
  • asnCountry (string | undefined): the country code of the AS of the client IP address. This is the administrative country of the AS, not necessarily the country of the client IP address.

IP type

The service field may be undefined, but you can use the hasService() method to check the availability. Using this method will also refine the type to remove the need for null-ish checks.

The following are available on the Free and Pro plans, and above:

  • hasService() (bool): whether the service field is available.
  • service (string | undefined): the name of the service associated with the IP address—e.g. Apple Private Relay.
  • isHosting() (bool): returns whether the IP address of the client is owned by a hosting provider. Requests originating from a hosting provider IP significantly increase the likelihood that this is an automated client.
  • isVpn() (bool): returns whether the IP address of the client is owned by a VPN provider. Many people use VPNs for privacy or work purposes, so by itself this is not an indicator of the client being automated. However, it does increase the risk score of the client and depending on your use case it may be a characteristic you wish to restrict.
  • isProxy() (bool): returns whether the IP address of the client is owned by a proxy provider. Similar to isVpn(), but proxies are more likely to involve automated traffic.
  • isTor() (bool): returns whether the IP address of the client is known to be part of the Tor network. As with isVpn(), there are legitimate uses for hiding your identity through Tor, however it is also often a way to hide the origin of malicious traffic.
  • isRelay() (bool): returns whether the IP address of the client is owned by a relay service. The most common example is Apple iCloud Relay, which indicates the client is less likely to be automated because Apple requires a paid subscription linked to an Apple account in good standing.

Error handling

Arcjet is designed to fail open so that a service issue or misconfiguration does not block all requests. The SDK will also time out and fail open after 1000ms when NODE_ENV or ARCJET_ENV is development and 500ms otherwise. However, in most cases, the response time will be less than 20-30ms.

If there is an error condition, Arcjet will return an ERROR conclusion.

import arcjet, { slidingWindow } from "@arcjet/bun";
import { env } from "bun";


const aj = arcjet({
  key: env.ARCJET_KEY!,
  // Tracking by ip.src is the default if not specified
  //characteristics: ["ip.src"],
  rules: [
    slidingWindow({
      mode: "LIVE",
      interval: "1h",
      max: 60,
    }),
  ],
});


export default {
  port: 3000,
  fetch: aj.handler(async (req) => {
    const decision = await aj.protect(req);


    if (decision.isErrored()) {
      // Fail open by logging the error and continuing
      console.warn("Arcjet error", decision.reason.message);
      // You could also fail closed here for very sensitive routes
      //return new Response("Service unavailable", { status: 503 });
    }


    if (decision.isDenied()) {
      return new Response("Forbidden", { status: 403 });
    }


    return new Response("Hello world");
  }),
};

Ad hoc rules

Sometimes it is useful to add additional protection via a rule based on the logic in your handler; however, you usually want to inherit the rules, cache, and other configuration from our primary SDK. This can be achieved using the withRule function which accepts an ad-hoc rule and can be chained to add multiple rules. It returns an augmented client with the specialized protect function.

import arcjet, { detectBot, shield } from "@arcjet/bun";
import { env } from "bun";


const aj = arcjet({
  key: env.ARCJET_KEY!,
  rules: [
    // Protect against common attacks with Arcjet Shield
    shield({
      mode: "LIVE", // will block requests. Use "DRY_RUN" to log only
    }),
  ],
});


function getClient(userId?: string) {
  if (userId) {
    return aj;
  } else {
    // Only apply bot detection to non-authenticated users
    return aj.withRule(
      detectBot({
        mode: "LIVE", // will block requests. Use "DRY_RUN" to log only
        allow: [], // "allow none" will block all detected bots
      }),
    );
  }
}


export default {
  port: 3000,
  fetch: aj.handler(async (req) => {
    // This userId is hard coded for the example, but this is where you would do a
    // session lookup and get the user ID.
    const userId = "totoro";


    const decision = await getClient(userId).protect(req);


    if (decision.isDenied()) {
      return new Response("Rate limited", { status: 429 });
    }


    return new Response("Hello world");
  }),
};

IP address detection

Arcjet will automatically detect the IP address of the client making the request based on the context provided. The implementation is open source in our @arcjet/ip package.

In development environments (NODE_ENV === "development" or ARCJET_ENV === "development"), we allow private/internal addresses so that the SDKs work correctly locally.

Client override

The default client can be overridden. If no client is specified, a default one will be used. Generally you should not need to provide a client - the Arcjet SDK will automatically handle this for you.

import arcjet, { slidingWindow, createRemoteClient } from "@arcjet/bun";
import { baseUrl } from "@arcjet/env";


const client = createRemoteClient({
  baseUrl: baseUrl(Bun.env),
  // timeout is the maximum time to wait for a response from the server. It
  // defaults to 1000ms when NODE_ENV or ARCJET_ENV is "development" and 500ms
  // otherwise. This is a conservative limit to fail open by default. In most
  // cases, the response time will be <20-30ms.
  timeout: 500,
});
import { env } from "bun";


const aj = arcjet({
  key: env.ARCJET_KEY!,
  // Tracking by ip.src is the default if not specified
  //characteristics: ["ip.src"],
  rules: [
    slidingWindow({
      mode: "LIVE",
      interval: "1h",
      max: 6,
    }),
  ],
  client,
});


export default {
  port: 3000,
  fetch: aj.handler(async (req) => {
    const decision = await aj.protect(req);


    for (const result of decision.results) {
      console.log("Rule Result", result);


      if (result.reason.isRateLimit()) {
        console.log("Rate limit rule", result);
      }


      if (result.reason.isBot()) {
        console.log("Bot protection rule", result);
      }
    }


    return new Response("Hello world");
  }),
};

Version support

Bun

As Bun is under active development, Arcjet aims to support Bun releases since v1.1.27.

When a Bun version goes end of life, we will bump the major version of the Arcjet SDK. Technical support is provided for the current major version of the Arcjet SDK for all users and for the current and previous major versions for paid users. We will provide security fixes for the current and previous major SDK versions.

Discussion