Discord Captcha Verification: How It Works and Recommended Settings
Discord captcha verification makes new members pass a challenge before posting, such as a distorted code answered with /verify. Discord's Raid Protection asks for a CAPTCHA only for an hour after it detects a join raid. A challenge on every join needs a bot. Security Bot offers an image captcha, a no-click web check, and Adaptive Verification, which picks between them per member. All are free.
Why a captcha and not just a verification level
Discord's verification level checks the account: verified email, account age, time in the server, or a verified phone. An account that is old enough and has a verified email passes. A verified phone number skips the level entirely.
A captcha checks the moment of joining instead: something has to solve a challenge before the account can post. That raises the cost of every automated account, and accounts that never solve it stay stuck in the verify channel. Captcha-solving services exist, so treat it as a cost for attackers, not a wall.
The cost for real members is a few seconds of friction. The settings below keep it low.
Image captcha vs web check
| Captcha code (image, in Discord) | Web check (browser, no clicks) | |
|---|---|---|
| Where it happens | In the verify channel: press Verify, get a code image only you can see | On a personal securitybot.gg link from the Verify button or a DM, with no sign-in |
| What the member does | Runs /verify captcha: with the code from the image | Keeps the page open for about 5 seconds |
| Against automation | A script has to read the distorted code, and /verify only accepts a code the bot issued | Denies browsers that show automation, though a well-hidden bot can pass. Can also block VPNs and alt accounts |
| Friction | Low, no leaving Discord | Low, but opens a browser |
| Accessibility | Visual only; keep codes short and keep a manual fallback | Nothing to read or solve |
| Typical fit | Gaming and hobby communities | Crypto, NFT, marketplaces, servers that attract alts |
Both types use the same Unverified role and verify channel, so you can switch between them from the dashboard without rebuilding anything. Adaptive Verification uses both. It sends each member straight in, to the image captcha, or to the web check, based on risk.
Recommended captcha settings
Security Bot's image captcha has six settings under Captcha settings on the Verification page. These starting values keep real members moving:
- Code length: 5 or 6 characters. The range is 4 to 8. Longer codes are harder to read and type on a phone.
- Code timeout: 2 minutes. Long enough to find the reply on mobile. Short enough that a stuck code expires and the next Verify press gives a new one. The options run from 30 seconds to 5 minutes.
- Extra lines: 1. Draws lines through the code so it is harder to read automatically, without making it unreadable. Raise it if accounts that look automated keep getting through.
- Include numbers: on. Mixes digits into the letters.
- Case sensitive: off to start. Turn it on if you want each code to be harder to guess.
- Code color: pick one that stands out clearly against the image background, and check it on a phone.
Under attack?
Adaptive Verification tightens on its own when joins spike, so more members get a check. On Captcha code alone, consider switching to the web check with VPN and alt blocking during an active raid. It raises the cost of each fake account, but it does not make one impossible.
How members verify
With Captcha code, the member joins and sees only the verify channel. They press Verify on Security Bot's message and get a code image in a reply only they can see. Then they run /verify captcha: with the code. A wrong code gets a retry, and an expired code is replaced with a new one.
With the web check, Security Bot tries to DM the member a personal link when they join, and pressing Verify gives them one too. The link opens securitybot.gg with no sign-in, and the page finishes on its own in about 5 seconds. Each link lasts 30 minutes.
With Adaptive Verification, pressing Verify gives the member whichever check the bot picked for them. Low-risk members never see the gate. In every case, Security Bot removes the Unverified role and adds the Verified role if you set one.
Enable verification and create the role and channel
Run
/verification enable,/verification setup_role, and/verification setup_channel, or use the Verification page in the dashboard. On Premium the commands reply with a link to that page.Command:
/verification enableChoose the verification type
Pick Adaptive to let low-risk members straight in and check the rest, or Captcha code to give everyone the image captcha. Web check swaps the captcha for a 5-second browser check.
Tune the image captcha
Under Captcha settings, pick a code length of 5 or 6 characters, a 2-minute timeout, and one extra line. Leave case sensitivity off to start.
Tell members what to expect
Security Bot's prompt in the verify channel explains the Verify button. Add one line saying who to ping if the code does not work.
Keep a manual fallback
The server owner or an Administrator ranked above Security Bot can approve a member who cannot solve the captcha.
Command:
/verification verify
Handling members who fail the captcha
- Expired code: the member presses Verify again for a fresh code. Keep the timeout at 2 minutes so this is rare.
- Cannot read the image: the owner or an Administrator ranked above Security Bot approves them with
/verification verify. Say in the channel who to ping. - Repeated failures from one account: that can be an automated account. Leave it in the verify channel. It cannot post anywhere else. In Adaptive mode, three wrong codes in 24 hours move it to the web check.
- Accessibility needs: the web check has nothing to read or solve. Or switch to one-click verification and lean on Discord's Raid Protection and, on Premium, anti-raid scoring instead.
Captcha settings by server type
| Server | Type | Image captcha settings | Pair with |
|---|---|---|---|
| Small community, under 1,000 members | Adaptive on Relaxed, or one-click | 5 characters, 2-minute timeout, case sensitivity off | Discord verification level Medium |
| Public gaming or fan server, 1,000 to 20,000 | Adaptive on Balanced | 6 characters, 2-minute timeout, 1 extra line, numbers on | Verification level High; anti-raid on Premium |
| Crypto, NFT, marketplace, or 20,000+ members | Adaptive on Strict, or Web check; VPN and alt blocking on | 6 characters, 1 extra line (Adaptive only) | Verification level High; Beast Mode; on Premium, anti-raid with a low score limit and Threat Link Guard |
The rule behind these is simple: the more money or reputation an account can extract from your server, the more each account should cost an attacker. The setup guide covers the rest of the bot.
Common captcha mistakes
- Maximum difficulty on day one. Eight characters, case sensitive, the most extra lines. Real members on phones fail, staff get flooded with requests, and the gate gets switched off. Start at 5 or 6 characters and tighten only if automated accounts get through.
- A 30-second timeout. The minimum is fine on desktop and hard for someone who joined from a phone notification. Use two minutes.
- `/verify` blocked in the verify channel. With Captcha code, members answer by running
/verifyin the channel. Security Bot's own setup allows Send Messages and Use Application Commands there. If you built the channel by hand, allow both for Unverified. Test with an alt before you announce the gate. - No instructions. Security Bot's prompt explains the button. Add one line saying who to ping if it fails, so members who get stuck do not just leave.
- Bot role below Unverified. Security Bot cannot remove a role above its own, so members pass the captcha and stay locked out. Move the bot's role up.
- Forgetting `/verification update_role`. Setup hides every existing channel from Unverified. But each channel you create later, of any type, needs the denial added, or unverified members can see it. The command adds it wherever it is missing.
Troubleshooting captcha verification
- The code never appears. The code is not posted in the channel. The member has to press Verify, and it arrives in a reply only they can see. If the channel itself looks empty, check that Unverified is allowed View Channel and Read Message History there.
- The code keeps expiring. The timeout is too short for your members. Raise it toward the 5-minute maximum, or switch to Adaptive or the web check if the problem persists.
- The web check link is invalid or expired. Links last 30 minutes, and each new link replaces the last. Have the member press Verify again for a fresh one.
- The web check says VPNs aren't allowed. VPN blocking is on. The member turns off their VPN or iCloud Private Relay and reloads, or you turn the block off.
- The web check says the device was already used. Alt blocking matched another account that verified from the same browser or device. If it is a shared family device, approve them with
/verification verify. - A member passed but still sees only `#verify`. Security Bot's role is below Unverified, so the removal failed. Fix the role order, then approve them with
/verification verify. - You want to switch types mid-raid. Change the type on the Verification page. The Unverified role and the channel stay as they are, so nothing needs rebuilding.
Passes are not written to Security Bot's log channels. Web check denials and possible alt accounts go to the security log. If a member insists they verified and nothing changed, check their roles. If they still hold Unverified, the removal failed. The verification docs list every option if you need more.
Premium captcha customization
The captcha itself is free. Premium adds a custom verification message and banner in the verify channel. Premium+ runs your own bot with a custom name, avatar, and status, so the captcha prompt comes from your server's own bot.
What a captcha does not protect against
A captcha shows that something solved a challenge at join time. It does not cover the rest of the attack surface, which is why Security Bot ships it next to other modules:
- Human-operated scam accounts. An aged account with a real person behind it solves the captcha and posts a scam link later. Discord's suspicious-link filter in Community servers and, on Premium, AutoMod with Threat Link Guard catch many of these links. Threat Link Guard deletes the link after it is posted and can time out, kick, or ban the sender.
- A verified member who starts flooding. On Premium, anti-spam counts the messages in each channel within a time window, then clears them and locks the channel.
- Join waves. Hundreds of accounts can still join and fill your verify channel. Discord's Raid Protection detects join raids. On Premium, anti-raid scores the wave, times out or bans the accounts that joined during it, and locks the channels you choose.
- Compromised staff. No verification setting stops an admin whose account was phished. That is anti-nuke.
Do this with Security Bot
Discord Verification Bot: Gate new members with Adaptive Verification, an image captcha, a no-click web check, one button, a passcode, or an allowlist.
Security Bot is a Discord verification bot that keeps new members behind an Unverified role until they pass. Adaptive Verification picks the check per member: low-risk accounts get straight in, others get an image captcha or a 5-second web check. When joins spike, it raises the bar on its own. Passing removes the role and can add your Verified role.
Frequently asked questions
Does Discord have a built-in captcha?
Partly. Discord's Raid Protection (Server Settings > Safety Setup > Raid Protection and CAPTCHA) asks new joiners to complete a CAPTCHA for an hour after it detects a join raid. For a challenge on every join, you need a bot such as Security Bot.
Is Security Bot's captcha free?
Yes. Adaptive Verification, image captcha, the web check, and one-click verification are all free. Premium adds a custom message and banner.
Which is safer, image captcha or the web check?
They stop different things. The image captcha makes a script read a distorted code. The web check denies detectable automation and can block VPNs and alt accounts. Adaptive uses both. None makes automation impossible, but each raises the cost of every fake account.
What captcha length should I use?
Five or six characters with one extra line and a two-minute timeout is a good start. Longer codes and more lines are harder for members on phones to read.
What if a real member cannot solve the captcha?
The server owner or an Administrator ranked above Security Bot can run /verification verify to approve them manually.
Can I run image captcha and the web check at the same time?
Yes, with Adaptive Verification. It sends each member straight in, to the image captcha, or to the web check, based on account age, history, and your join rate. You can switch types at any time without recreating the role or channel.
Does captcha verification work on mobile?
Yes. The code image arrives in a reply in the Discord app, and the web check opens securitybot.gg in the phone's browser. Keep codes at 5 or 6 characters so they are readable on small screens.
How this guide was made: drafted with AI assistance, then checked against Security Bot's source code and Discord's official help articles. Spotted something out of date? Tell us in the support server.