Skip to main content
Synced from an Obsidian vault

For graph and advanced features, download the full Intel Codex Vault and open it in Obsidian.

Discord OSINT SOP

Authorized OSINT use only. Investigation accounts must comply with Discord's Terms of Service and the platform's prohibitions on bulk scraping, automated account creation, self-bots, and coordinated inauthentic behavior. Review Legal & Ethics SOP and OPSEC Planning before any engagement. CSAM, threat-to-life DMs, or trafficking content trigger an immediate hard stop and routing per Sensitive Crime Escalation SOP.

Table of Contents​

  1. Access & Client Options
  2. Discovery & Search Methods
  3. Search Operators & Methods
  4. Repeatable Workflow
  5. Server & Channel Analysis
  6. User Profile Intelligence
  7. Investigation Recipes
  8. Collection & Evidence Integrity
  9. Advanced Techniques
  10. Pivoting & Cross-Platform Correlation
  11. Tools & Resources
  12. Risks & Limitations
  13. Quality Assurance & Verification
  14. Real-World Scenarios
  15. Emergency Procedures
  16. Legal & Ethical Considerations
  17. Related SOPs
  18. External Resources

1) Access & Client Options​

  • Web: https://discord.com/app (full feature parity with desktop, runs in browser)
  • Desktop apps: Windows / macOS / Linux clients with full feature access
  • Mobile apps: iOS / Android with most features (some admin actions desktop-only)
  • Official API: Discord HTTP + Gateway APIs documented at https://discord.com/developers/docs (bot accounts only; user-token automation is a ToS violation called "self-botting")
  • Bot integration: discord.py (Python), discord.js (Node), serenity (Rust), JDA (Java) — bot must be invited to a server by an admin via OAuth2 URL
  • Authorized export tooling: DiscordChatExporter (CLI/GUI) — see §8 for ToS framing
  • Third-party clients: BetterDiscord, Powercord/Replugged, and similar client mods are ToS violations and risk account termination — do not use on investigation accounts

2) Discovery & Search Methods​

Server Discovery​

  • Direct invites: discord.gg/<code> and discord.com/invite/<code> resolve to a server preview (server name, icon, approximate member count, approximate online count) without joining
  • Vanity URLs: discord.gg/&lt;vanity&gt; — vanity codes require Boost Level 3 to claim [verify 2026-04-27 — Discord has rotated boost-tier reward thresholds; current boost count for Level 3 may differ]; presence of a vanity is a weak signal of established/funded server
  • Server listing sites: Disboard (https://disboard.org), Disforge (https://disforge.com), Top.gg server directory (https://top.gg/servers), Discords.com (https://discords.com), DiscordHome (https://discordhome.com) — quality varies; treat as discovery surface, not authoritative inventory
  • Bot listing sites: Top.gg bot directory, Discord Bot List (https://discordbotlist.com) — useful for identifying bots seen in target server
  • Cross-platform mentions: Twitter/X bios, Reddit posts, Telegram channels, GitHub READMEs, YouTube descriptions frequently link to Discord invites
  • Invite-graph pivots: a single user posting an invite across 5 unrelated platforms is a coordination signal — capture all source posts before the invite is rotated or revoked

Invite Resolution Without Joining​

  • The unauthenticated endpoint https://discord.com/api/v10/invites/<code>?with_counts=true&with_expiration=true returns server name, ID, icon hash, splash hash, approximate_member_count, approximate_presence_count, expiration timestamp, inviter (if set), and the target channel — without joining and without an authenticated session [verify 2026-04-27 — endpoint version may have rotated]
  • Useful for documenting a server's existence and rough size before deciding whether to join with a sock-puppet account
  • Rate-limited per IP; do not script bulk resolution against this endpoint

Search Techniques​

  • Username enumeration: try common variations (underscores, numbers, leetspeak) — note Discord moved from username#discriminator (legacy) to globally unique handles in 2023; both forms may surface in old screenshots/logs
  • Server name search: server-listing sites support keyword search and tag filters (gaming, crypto, NFT, anime, etc.)
  • Channel structure inference: common naming patterns (#announcements, #rules, #general, #trades, #vendors, #leaks) correlate with server purpose
  • Bot fingerprinting: presence of MEE6, Dyno, Carl-bot, Tatsu indicates moderation maturity; Tip.cc, Collab.Land, Vulcan indicate crypto/Web3 focus

3) Search Operators & Methods​

Discord's in-app search supports a limited operator set scoped to the current server (DMs and group DMs have their own search). Key operators (verify current set against client UI — Discord rotates these):

  • from:@user — messages authored by a specific user
  • mentions:@user — messages mentioning a specific user
  • has:link / has:embed / has:file / has:video / has:image / has:sound — content type filter
  • before:YYYY-MM-DD / during:YYYY-MM-DD / after:YYYY-MM-DD — date range filter
  • in:#channel — restrict to a specific channel
  • pinned:true — pinned messages only
  • Operators stack: from:@vendor has:link in:#trades after:2024-01-01
  • Quote exact phrases with double quotes: "contract address"
  • Limitation: search results are capped (UI shows up to ~25 per page, with paginated continuation); for completeness, use authorized export

4) Repeatable Workflow​

  1. Target identification: server invite code, vanity URL, user IDs, bot IDs, suspected aliases
  2. Pre-join intel: resolve invite metadata (member counts, server ID, channel ID); decide whether joining is in scope
  3. Sock-puppet selection: choose investigation account appropriate for server profile (gaming, crypto, hobby) — see OPSEC Planning
  4. Server enumeration: capture full channel list, role list, emoji set, server boost level, server creation date (from server snowflake ID)
  5. Member-list discipline: large servers (10k+ members) only render online members in the sidebar by default; full member list requires scrolling, bot enumeration via authorized API, or [verify 2026-04-27 — Member List Update gateway opcode behavior]
  6. Lurking phase: passive observation, role-assignment to limit detectability (avoid roles that grant moderator visibility), no posting
  7. Content harvest: targeted message capture (screenshots + authorized export); preserve full chain (server → channel → message ID)
  8. User-ID resolution: extract snowflake IDs (Developer Mode → right-click → Copy ID) for every entity of interest
  9. Cross-server pivots: shared bots, shared moderators, shared invite-poster accounts → enumerate adjacent servers
  10. Documentation: timestamps in UTC, SHA-256 hashes for media, log entries in Collection Log

5) Server & Channel Analysis​

Server Intelligence​

  • Server ID: snowflake ID — extractable via Developer Mode → right-click server icon → Copy Server ID
  • Server creation date: derived from server snowflake (see §6 snowflake structure); useful for distinguishing established communities from rapidly stood-up scam infrastructure
  • Boost Level (0–3): indicates community investment; vanity URL and animated icon require Level 3
  • Verification Level: Discord's anti-spam tier (None / Low / Medium / High / Highest) — visible to members, affects who can post
  • Explicit Content Filter: server-wide image scanning setting
  • Server description / About: discoverable via server-listing sites and the Discovery feature for Community-enabled servers
  • Welcome Screen / Onboarding: Community-feature servers expose channel hints and rule summaries to new members — capture before joining if possible

Channel Intelligence​

  • Channel types: Text, Voice, Announcement, Stage, Forum, Media, Thread (parent-child), Category (organizational only)
  • Channel ID: snowflake — Developer Mode → right-click channel → Copy Channel ID
  • Channel creation date: derived from snowflake; channels created during a single short window suggest scripted server setup (template, bot, or copied server)
  • Permission overwrites: each channel can override server-wide role permissions (visible to members with View Channel)
  • Slowmode / NSFW flag / Age-restricted: channel metadata visible in channel settings preview
  • Pinned messages: often contain rules, vendor menus, contract addresses, or admin contact info — capture early
  • Forum channels (post-2022): each post is a thread with tags; search behaves differently from text channels

Role & Permission Mapping​

  • Role list: Server Settings → Roles (visible to all members in most servers); roles often encode hierarchy (@Owner, @Admin, @Mod, @Vendor, @Verified, @Booster)
  • Permission audit: roles with Administrator, Manage Server, Manage Channels, Ban Members, Kick Members, Manage Roles, Manage Webhooks are operationally significant
  • Hoisted roles: Display role members separately — visible in member sidebar, useful for quick admin enumeration
  • Role color: cosmetic but consistent within a server hierarchy

Audit Log Limits​

  • Audit Log access: requires View Audit Log permission (typically admin/mod only); the analyst account in a lurker role will not see audit logs
  • Retention: Discord retains audit logs for a limited window (commonly cited as 45–90 days depending on source) [verify 2026-04-27 — confirm current retention against Discord's published documentation before relying on a specific number in a report]
  • Coverage: audit log captures structural changes (channel create/delete, role create/delete, ban, kick, member role change, webhook create) — does not capture message content

6) User Profile Intelligence​

Profile Analysis​

  • Global Name (display name): user-chosen, server-overridable, may not match handle
  • Username (handle): globally unique, lowercase, post-2023 migration replaces the legacy username#discriminator form
  • Discriminator (legacy): #1234 four-digit suffix — phased out in 2023 but still appears in archived screenshots / logs / older bots
  • Server Nickname: per-server alias, set by the user or by mods; differs from global name
  • Bio / About Me: free-text profile field, supports limited markdown and links
  • Pronouns: profile field added 2023
  • Profile picture (avatar): server-overridable; reverse-image search candidate, often reused across platforms
  • Banner: Nitro-only profile field
  • Connections: linked accounts (Twitch, YouTube, Steam, Spotify, GitHub, Twitter/X, Reddit, eBay, PlayStation, Xbox, etc.) — visible per the user's privacy settings; high-value pivot when present
  • Mutual servers / Mutual friends: visible to anyone who shares at least one server with the target (privacy setting)
  • Activity / Status: "Playing", "Listening to Spotify", "Streaming on Twitch" — leaks platform usage and timezone
  • Custom Status: free-text status, often contains contact info (Telegram handle, email, Tox ID)

Discord Snowflake Structure (timestamp extraction)​

Discord IDs (user, server, channel, message, role, webhook, etc.) are snowflake IDs: 64-bit integers whose top bits encode a millisecond-precision creation timestamp. Contrast with Telegram, where user IDs are monotonically-increasing integers without an embedded timestamp (see Telegram SOP §6).

Extraction formula:

unix_ms = (snowflake_id >> 22) + 1420070400000

Where 1420070400000 is Discord's epoch (2015-01-01T00:00:00.000Z) in milliseconds since the Unix epoch. The result is a millisecond Unix timestamp; divide by 1000 for seconds.

Python example:

import datetime

DISCORD_EPOCH_MS = 1_420_070_400_000

def snowflake_to_datetime(snowflake: int) -> datetime.datetime:
    """Convert a Discord snowflake ID to its UTC creation timestamp."""
    unix_ms = (snowflake >> 22) + DISCORD_EPOCH_MS
    return datetime.datetime.fromtimestamp(unix_ms / 1000, tz=datetime.timezone.utc)

# Example: a user ID
user_id = 175928847299117063  # any valid snowflake
print(snowflake_to_datetime(user_id))  # -> 2016-04-30 11:18:25.796000+00:00

One-liner (any shell with python3):

python3 -c 'import sys, datetime; s=int(sys.argv[1]); print(datetime.datetime.fromtimestamp(((s>>22)+1420070400000)/1000, tz=datetime.timezone.utc))' 175928847299117063

Operational uses:

  • Account creation date from a user's snowflake → flags freshly-created accounts in a coordinated server
  • Server creation date from a server's snowflake → distinguishes established communities from rapid scam stand-up
  • Channel creation date from a channel's snowflake → channels all created within a short window suggest scripted setup (server template / bot / copy)
  • Message creation date from a message's snowflake → independent of any client-displayed timestamp; useful when a screenshot's timestamp is suspect
  • Webhook creation date → see §9 webhook discovery

The 22-bit shift discards: 5 bits internal worker ID, 5 bits process ID, 12 bits sequence counter — none of which are operationally useful for OSINT.

Behavioral Analysis​

  • Posting cadence: tracked across captured messages; identifies timezone clusters
  • Cross-server presence: mutual-servers UI shows shared servers (subject to privacy setting); valuable for identifying community overlap
  • Reaction patterns: who reacts to what — note that reactions are visible and tracked
  • Voice activity: who joins which voice channels and when (visible in voice channel member list while joined)
  • Bot interactions: which slash commands a target invokes, which bots they have authorized via OAuth
  • Username changes: Discord does not publish username history; bots like Vulcan and some moderation bots track and expose history within their own server (verify per server)

7) Investigation Recipes​

Ransomware Affiliate Coordination​

Objective: Document affiliate / partner coordination on Discord servers operated by or adjacent to a ransomware group.

Method:

  1. Identify the server via leak-site links, dark-web mirror references, or actor-profile pivots from Darkweb Investigation
  2. Resolve invite metadata (server ID, member count) before joining
  3. Join with a sock-puppet account meeting any vetting requirements; do not engage
  4. Capture server structure: channel list, role list, pinned messages, rules, vendor verification process
  5. Document affiliate-recruitment messages, payment terms, victim-targeting criteria, share of proceeds
  6. Extract crypto wallet addresses from payment instructions; route to Blockchain Investigation and Financial AML OSINT
  7. Map admin / partner accounts; pivot to other servers via mutual-servers and shared-bot heuristics
  8. Preserve evidence with timestamps, snowflake IDs, and SHA-256 hashes per Collection Log
  9. Escalate per engagement RoE — if active threat-to-life or imminent victim impact, route per Sensitive Crime Escalation SOP

No-KYC Crypto Swap / OTC Server​

Objective: Document no-KYC swap or OTC server acting as a money-laundering layer.

Method:

  1. Locate via server-listing sites (search: "OTC", "swap", "instant exchange") and Telegram cross-references
  2. Document fee structure, supported chains, minimum/maximum amounts, payment workflow
  3. Capture deposit-address rotation patterns and bot interaction transcripts
  4. Pivot identified deposit addresses to Blockchain Investigation and Mixer & Privacy-Pool Tracing via the Financial OSINT path
  5. Capture admin / operator handles for cross-platform correlation (Telegram, X, dark-web forums)

Scam Community (Pump-and-Dump / Rug-Pull / Pig-Butchering / Fake Airdrop)​

Objective: Document a coordinated scam community and its victim funnel.

Method:

  1. Identify entry-point invites posted on Twitter/X, TikTok, Telegram, or via paid Disboard bumps
  2. Capture the funnel: landing channel → "verification" bot → role assignment → "alpha" channel
  3. Document the contract address(es), trade-call timing, and admin/influencer rotation
  4. Extract victim-report messages, complaint patterns, "where's my money" channels
  5. Cross-reference contract addresses with token-launch tooling (pump.fun, Solana SPL launchers) and DEX trade history
  6. Preserve admin-DM screenshots only when consensual or under explicit authorization (sock-puppet engaging in active conversation crosses into pretexting — see §16)

Credential-Stealer Logs Market​

Objective: Document Discord-based markets selling stealer logs (RedLine, Raccoon, LummaC2, StealC, etc.) without acquiring or analyzing the logs themselves.

Method:

  1. Identify markets via dark-web forum cross-posts, Telegram channel links, and Discord-listing-site search
  2. Document seller pricing, log-volume claims, geo/category filters offered, sample-set policy
  3. Capture seller handles, admin handles, payment wallets, bot-mediated checkout flow
  4. Do not download logs. Stealer logs contain real-victim PII and credentials; possession may violate computer-misuse statutes and engagement RoE. Route stealer-malware analysis to Malware Analysis SOP
  5. If logs include child-victim data or imminent-threat indicators, hard-stop and escalate per Sensitive Crime Escalation SOP

Geographic / Community Intelligence​

Objective: Local-incident reporting and community intel via city/region servers.

Method:

  1. Locate local servers via Disboard tag filter (city / region / language)
  2. Search incident keywords: protest, blackout, shooting, evacuation, scam (in local language)
  3. Cross-reference timestamps with corroborating sources (local news, X, Reddit)
  4. Geolocate posted media via Image & Video OSINT
  5. Verify with multiple sources before attribution

8) Collection & Evidence Integrity​

Authorized Export​

  • Discord's first-party data export (Settings → Privacy & Safety → Request All My Data) returns ZIP of analyst's own messages / servers / connections — useful for investigator-account record-keeping, not third-party-server collection
  • DiscordChatExporter (https://github.com/Tyrrrz/DiscordChatExporter): CLI/GUI exports a channel to HTML / JSON / plaintext / CSV. ToS framing:
    • Self-owned data: clearly permitted (analyst's own DMs, analyst-owned servers)
    • Authorized customer-server engagement: permitted under the server-owner's authority and within engagement scope
    • Third-party server (analyst joined as lurker): ToS gray area — Discord ToS prohibits scraping and data-mining; route the question to Legal & Ethics SOP before automating, document the legal basis in Collection Log
    • DiscordChatExporter operates against a user token by default (which is itself self-bot territory) or a bot token (requires bot to be present in the server, which requires admin-side OAuth). Use the bot path when the engagement permits it.
  • Authorized bot collection: discord.py / discord.js bots invited via OAuth2 with Read Message History and the necessary intents (Server Members Intent, Message Content Intent — both privileged, require Discord-side verification once the bot crosses the verification threshold) [verify 2026-04-27 — exact server-count threshold for required privileged-intent verification]. Cleaner audit trail than user-token tooling.

Manual Capture​

  • Screenshot capture: include channel name, server name, timestamp visible (or hover-revealed full timestamp); capture both compact and cozy modes when message structure matters
  • Message ID copy: Developer Mode → right-click message → Copy Message ID (snowflake captures exact creation time independent of client display)
  • Message link copy: right-click → Copy Message Link → returns https://discord.com/channels/&lt;server_id&gt;/&lt;channel_id&gt;/&lt;message_id&gt; — the canonical reference form, even though it requires server membership to resolve

File Organization​

/Evidence/{{case_id}}/Discord/
├── YYYYMMDD-HHMM/
│   ├── servers/
│   │   ├── &lt;server_id&gt;_metadata.json     # name, ID, owner, created_at, channel list, role list
│   │   ├── &lt;server_id&gt;_channels/
│   │   │   ├── &lt;channel_id&gt;_&lt;channel_name&gt;.html   # DiscordChatExporter HTML
│   │   │   └── &lt;channel_id&gt;_&lt;channel_name&gt;.json   # DiscordChatExporter JSON
│   │   └── &lt;server_id&gt;_screenshots/
│   ├── users/
│   │   └── &lt;user_id&gt;_&lt;handle&gt;_profile.png
│   ├── bots/
│   │   └── &lt;bot_id&gt;_&lt;handle&gt;_oauth_scopes.txt
│   ├── webhooks/
│   │   └── &lt;webhook_id&gt;_metadata.json
│   ├── media/
│   │   ├── images/
│   │   ├── videos/
│   │   └── attachments/
│   └── SHA256SUMS

Hashing & Verification​

  • SHA-256 every exported file and downloaded attachment immediately after capture
  • Store hashes in SHA256SUMS (one per capture batch)
  • Record snowflake IDs alongside human-readable names in the metadata file (names change; IDs are immutable)
  • Reference all captures in Collection Log with capture timestamp (UTC), snowflake ID, server / channel / message reference, and analyst account used

9) Advanced Techniques​

Webhook Discovery & Triage​

Discord webhooks have the URL form:

https://discord.com/api/webhooks/&lt;webhook_id&gt;/&lt;webhook_token&gt;
  • The &lt;webhook_id&gt; is a snowflake; the &lt;webhook_token&gt; is a long opaque secret — anyone holding the URL can post to the channel until the webhook is rotated/deleted
  • Webhooks frequently leak in:
    • Public GitHub commits (CI configs, .env files, JS bundles)
    • Public S3 / GCS buckets (config drift)
    • Mobile app reversed binaries
    • JavaScript files served from compromised sites
  • Do not POST to a third-party webhook URL — this is unauthorized access to a target system, regardless of how publicly the URL was exposed. Capture the URL, GET its metadata via GET https://discord.com/api/webhooks/&lt;id&gt;/&lt;token&gt; only when authorized in scope (this is read-only and returns webhook name, channel ID, server ID, avatar) [verify 2026-04-27]
  • Webhook leaks discovered during bug-bounty engagements: route disclosure through Bug Bounty Methodology SOP and report to the program; do not validate by posting

Bot Scope & Slash-Command Surface Review​

  • Bots in a server expose their slash commands via the / autocomplete UI; capture the full command list as part of server enumeration
  • Bot OAuth scopes can be inspected via the Authorized Apps page (Settings → Authorized Apps) for the analyst's account; for a target's bot, scopes can be inferred from the bot's invite URL: client_id=&lt;bot_id&gt;&scope=bot+applications.commands&permissions=&lt;permission_int&gt;
  • Decode the permissions integer (Discord publishes the bit table at https://discord.com/developers/docs/topics/permissions) to enumerate exactly what the bot can do — Administrator (bit 3) or Manage Server (bit 5) on a third-party bot is a strong supply-chain signal
  • High-risk permission combinations on a community bot:
    • Manage Webhooks + Manage Server → bot can pivot to webhook-based exfiltration
    • Manage Roles + position above @everyone → bot can self-elevate trusted users
    • Read Message History + Send Messages + privileged Message Content Intent → bot can mass-DM, scrape, or relay content out of band

Cross-Server Moderator / Bot Overlap​

  • Mutual-server enumeration on a confirmed admin account often reveals the operator's other servers
  • Shared custom bots (especially uncommon or self-hosted) across multiple servers indicate operational overlap
  • Reaction-emoji custom emoji are server-bound; a custom emoji from server A appearing as a reaction in server B reveals that the reactor has Nitro and is a member of server A

Voice & Stage Channel Surveillance Discipline​

  • Joining a voice channel announces presence to every member with View Channel permission on that voice channel — there is no silent observe mode
  • Stage Channels (audience-style): joining as audience does not announce as prominently as voice-channel join, but the audience list is visible to speakers and moderators
  • For evidence collection from voice content, prefer authorized recording with explicit operator consent or law-enforcement legal process; surreptitious recording violates wiretap statutes in many jurisdictions — consult Legal & Ethics SOP

Deleted Content​

  • Discord does not retroactively expose deleted messages
  • Continuous capture (authorized bot logging every message, or active analyst with notification cache) is the only OSINT path
  • Some moderation bots (Vulcan, Dyno, Carl-bot) keep a deletion log within their own server scope — visible only to staff with the relevant permission

10) Pivoting & Cross-Platform Correlation​

Profile Pivots​

  • Bio / About Me links → personal sites, X, Telegram, GitHub, Linktree
  • Connections panel → linked Twitch / YouTube / Steam / GitHub / Spotify / X / Reddit / PlayStation / Xbox / eBay accounts (visibility per user privacy)
  • Username → cross-platform handle search via Sherlock, Maigret, WhatsMyName (Entity Dossier consolidation)
  • Avatar → reverse-image search (Google Lens, Yandex, TinEye); avatars are reused across platforms more than display names
  • Custom Status → free-text often contains Telegram @handle, Tox ID, Session ID, or email
  • Snowflake ID → account creation date narrows the alias-overlap search

Content Pivots​

  • Shared URLs → website infrastructure via Web/DNS/WHOIS OSINT
  • Cryptocurrency addresses in messages → Blockchain Investigation / Financial AML OSINT
  • Smart-contract addresses in trade-call channels → contract source review via block explorer; route audit-style review to existing audit SOPs
  • Leaked API keys / cloud credentials posted in chat (AWS / GCP / Azure keys, Discord bot tokens, OpenAI keys, GitHub PATs are common in screenshots) → Cloud Pentesting SOP for authorized validation, or report to the affected vendor
  • Media files → reverse search, EXIF stripping behavior (Discord generally strips most EXIF from images on upload, but documents / archives / video containers retain their metadata) [verify 2026-04-27 — confirm current upload-pipeline behavior; Discord has changed this several times] — Image & Video OSINT
  • Document metadata → author, organization, file path artifacts in PDFs / Office docs

Network Pivots​

  • Mutual servers between two accounts → community overlap
  • Shared admin set across multiple servers → operator network
  • Shared bots (especially self-hosted or unlisted) → infrastructure overlap
  • Invite-poster account posting the same invite across X, Reddit, Telegram, TikTok → coordinated promotion
  • Reaction overlap: same set of reactor accounts on different servers' announcement posts → astroturf signal
  • Comparative messaging-platform pivot — Telegram channel linked from Discord server (or vice versa) is a common pattern; cross-reference Telegram SOP for the comparative channel/group methodology

11) Tools & Resources​

ToolPurposePlatformAccess
Discord DesktopPrimary investigation client (enable Developer Mode in Settings)Desktophttps://discord.com/download
DiscordChatExporterAuthorized channel/DM export (HTML/JSON/CSV/plaintext)CLI / GUIhttps://github.com/Tyrrrz/DiscordChatExporter
discord.pyPython bot library for authorized monitoringPythonpip install discord.py
discord.jsNode.js bot library for authorized monitoringNodenpm install discord.js
DisboardServer discovery & tag-filter searchWebhttps://disboard.org
DisforgeServer discovery directoryWebhttps://disforge.com
Top.ggServer + bot directoryWebhttps://top.gg
Discords.comServer directory (formerly discord.me)Webhttps://discords.com
Snowflake decoderConvert snowflake → UTC creation timestampWeb / CLIpython3 -c '...' (see §6) or https://snowsta.mp [verify 2026-04-27 — third-party site availability]
Sherlock / MaigretCross-platform username enumerationCLI / Pythonpip install maigret / Sherlock GitHub
WhatsMyNameUsername presence across hundreds of sitesWeb / CLIhttps://whatsmyname.app
Nuclei (osint tags)Templated username enumerationCLInuclei -tags osint -var user=&lt;handle&gt;
Discord API docsOfficial endpoint and gateway referenceDocshttps://discord.com/developers/docs
IntelXDiscord token / webhook leak searchWebhttps://intelx.io
GrepApp / GitHub searchCode-search for leaked Discord webhooks / tokensWebhttps://grep.app, https://github.com/search

API Usage Example (authorized bot)​

# discord.py - capture authorized channel history (analyst-side bot, invited via OAuth)
import os
import discord

intents = discord.Intents.default()
intents.message_content = True  # privileged intent; enable in Developer Portal

client = discord.Client(intents=intents)

@client.event
async def on_ready():
    target_channel_id = int(os.environ["DISCORD_CHANNEL_ID"])
    channel = client.get_channel(target_channel_id)
    async for msg in channel.history(limit=None, oldest_first=True):
        print(f"{msg.created_at.isoformat()} | {msg.id} | {msg.author} | {msg.content}")
        for att in msg.attachments:
            print(f"  attachment: {att.url} ({att.size} bytes)")
    await client.close()

client.run(os.environ["DISCORD_BOT_TOKEN"])

Snowflake Decode One-Liner​

# Decode a snowflake to UTC datetime
python3 -c 'import sys, datetime; s=int(sys.argv[1]); print(datetime.datetime.fromtimestamp(((s>>22)+1420070400000)/1000, tz=datetime.timezone.utc))' 175928847299117063

12) Risks & Limitations​

  • ToS prohibition on scraping & self-bots: user-token automation, mass DMs, automated account creation, and scraping are explicitly prohibited; investigation accounts that violate ToS risk ban and evidence inadmissibility
  • Coordinated inauthentic behavior detection: Discord aggressively bans linked sock-puppet accounts (shared payment, shared IP, shared device fingerprint, similar registration cadence)
  • Voice-channel join announces presence to all members with View Channel — there is no silent observation mode
  • DM acceptance signal: opening a DM may notify the recipient depending on their privacy settings (server-shared default, friends-only, off); consider this before initiating
  • Reaction-emoji visibility: reactions are public to the channel and persistent until removed — a reaction is evidentially equivalent to a post
  • Friend request signal: sending a friend request notifies the target
  • Profile / Connections privacy: targets can hide mutual servers, mutual friends, connections, and "Now Playing" — absence of data is not absence of activity
  • Phone / email obscured: Discord does not surface email or phone to other users; data-broker correlation requires legal process or breach data
  • Account verification phone requirement: high-risk-flagged servers may require verified phone on the joining account; investigation phones must be isolated from analyst identity per OPSEC Planning
  • Rate limiting: REST endpoints rate-limit per route per token; Gateway has connection identify limits — bulk enumeration scripts will be throttled or terminated
  • Content deletion: messages, channels, servers, and accounts can disappear without warning; capture early
  • Server template / mass-deployment churn: many scam servers are stood up from a template, banned, and re-deployed within hours — sustained tracking requires automation under engagement scope
  • Malware risk: attachments in criminal servers may contain malicious payloads; never open on the analyst host — route to Malware Analysis SOP sandbox

13) Quality Assurance & Verification​

Source Verification​

  • Snowflake-derived timestamps are authoritative for creation time of the entity (server, channel, message, user) — use them to validate any user-displayed timestamp
  • Server authenticity: Community-feature servers expose verified status for partners; "verification" claimed in a server's About text is not the same as Discord's verification — confirm via the badge on the server icon
  • Bot identity: a bot's client_id (snowflake) is fixed; impostor bots use similar names but new IDs — capture and compare
  • Cross-source corroboration: cross-reference claims with Telegram channels, X profiles, dark-web forum posts, and on-chain activity before attribution

Evidence Validation​

  • Snowflake IDs alongside every captured artifact (server, channel, message, user)
  • Both human-readable name and ID in collection log — names mutate, IDs do not
  • Capture context: include channel name, server name, server ID in every screenshot (or in the filename)
  • Hash chain: SHA-256 immediately after capture; entries in Collection Log
  • Sock-puppet account documentation: which investigation account observed each artifact, with the account's own snowflake recorded once per case

14) Real-World Scenarios​

Scenario 1: Stealer-Logs Market Disruption Support​

Situation: Analyst-tracked Discord server selling RedLine and LummaC2 stealer logs at $5–$50 per log, organized by geo and credential type.

Approach:

  1. Confirm jurisdiction and engagement RoE before joining; route through legal review
  2. Resolve invite metadata and document server ID, member counts, channel structure pre-join
  3. Join with a sock-puppet account that fits the community's vetting (often only requires reading rules + reacting to a verification emoji)
  4. Document seller pricing, stock claims, sample-policy, supported payment methods, OPSEC instructions to buyers, refund policy
  5. Capture seller and admin handles + snowflake IDs; pivot via mutual servers to identify the operator's network
  6. Extract crypto wallet addresses; route to Blockchain Investigation
  7. Do not download any logs. Stealer logs contain real-victim PII; possession may itself violate computer-misuse statutes
  8. Compile evidence package (screenshots, exports, hashes, snowflake IDs, wallet pivots) per Reporting & Disclosure SOP
  9. Hand off to law enforcement or the affected vendor security team

Outcome shape: Evidence package supports takedown coordination and victim-notification efforts; sustained tracking requires authorized bot collection, not user-token automation.

Scenario 2: Pump-and-Dump / Scam Token Community​

Situation: A 60k-member Discord server promoting "alpha calls" on Solana SPL tokens, with admin-rotated wallet addresses for "VIP access" payment.

Approach:

  1. Map the funnel: invite source (X promoted post / TikTok) → landing channel → verification bot → role-gated alpha channel
  2. Capture announcement-channel call timing relative to on-chain trade times (admins front-running their own calls is the dominant pattern)
  3. Document the contract addresses called and their launch / liquidity events via on-chain investigation
  4. Capture admin handles, snowflake IDs, Connections-panel data; pivot to other servers via mutual-servers
  5. Cross-reference launch wallets and recipient wallets to identify the operator's address cluster
  6. Document victim complaint patterns in #support / #general channels (often deleted by mods within hours — capture early)
  7. Package per Reporting & Disclosure SOP; consider responsible disclosure to the chain's compliance contacts and to relevant exchanges (deposit-address blocking)

Outcome shape: Evidence supports civil recovery and CFTC / SEC complaints in US scope; cross-reference with the financial AML OSINT path for tracing.

Scenario 3: Webhook Leak Bug-Bounty Triage​

Situation: During an authorized bug-bounty engagement, a public S3 bucket exposes a config.json containing a Discord webhook URL used by the customer's CI pipeline for build-failure notifications.

Approach:

  1. Do not POST to the webhook URL — that is unauthorized access to a customer system
  2. GET the webhook metadata via GET https://discord.com/api/webhooks/&lt;id&gt;/&lt;token&gt; to confirm ownership (returns webhook name, server ID, channel ID) — only if read-only metadata access is in scope
  3. Document the leak with: source URL of the leaked config, capture timestamp, file SHA-256, webhook ID (snowflake → creation timestamp), server ID, channel ID
  4. Report through the bug-bounty program's official channel per Bug Bounty Methodology SOP
  5. Recommend remediation: rotate webhook (delete + recreate), audit S3 bucket policy, enable secret-scanning in CI
  6. Do not publish the webhook URL in any report draft, screenshot, or external write-up — redact the token portion before any handoff

Outcome shape: Disclosure-grade evidence package; vendor rotates webhook within hours of report; finding may qualify for the program's information-disclosure tier.

15) Emergency Procedures​

Immediate Archival Triggers​

  • Server announces shutdown / migration / "moving to Telegram"
  • Admins begin mass channel deletion or role purges
  • Active threat-to-life content (violence planning, victim targeting)
  • CSAM or trafficking indicators
  • Law-enforcement preservation request directed at the engagement

Rapid Response Protocol​

  1. Immediate authorized export of in-scope channels via DiscordChatExporter (HTML + JSON, both formats — HTML for human review, JSON for machine pipeline)
  2. Screenshot server overview, channel list, role list, member sidebar (visible online members), pinned messages of every monitored channel
  3. Capture all snowflake IDs for server, channels, roles, key users — names mutate, IDs do not
  4. Download attachments at original quality; SHA-256 immediately
  5. External archival for any public-facing endpoints (server-listing-site profile, public bot pages) via Archive.org / archive.today
  6. Forward critical messages to a dedicated case "saved messages" workspace (analyst's own DMs) — preserves the message link form https://discord.com/channels/&lt;sid&gt;/&lt;cid&gt;/&lt;mid&gt; even if the source channel is later purged
  7. Hash everything and update Collection Log
  8. Notify engagement lead / client / law-enforcement liaison per RoE

Sensitive-Crime Hard Stop (CSAM, Trafficking, Imminent Threat-to-Life)​

  • Do not download the offending content
  • Capture URL + timestamp + snowflake ID + server / channel ID only
  • Do not analyze further
  • Route immediately per Sensitive Crime Intake & Escalation SOP
  • CSAM: report to NCMEC CyberTipline (US — https://report.cybertip.org), IWF (UK), INHOPE network national member elsewhere; preserve URL only
  • Imminent threat-to-life: local emergency services + Discord Trust & Safety (https://dis.gd/request) with emergency-disclosure framing

Escalation Triggers (see Sensitive Crime Escalation SOP)​

  • Immediate threats: terrorism, mass violence, specific threats to identifiable individuals
  • Child safety: CSAM, grooming, sextortion, trafficking
  • Active crimes: large-scale stealer markets with active victim-data flow, contract-killing-for-hire claims, weapons sales
  • National security: state-sponsored influence operations, espionage adjacency, CI targeting
  • Legal process: court orders, subpoenas, mutual-legal-assistance requests routed to the engagement lead

16) Legal & Ethical Considerations​

Discord OSINT sits at the intersection of platform ToS, sock-puppet pretexting law, and (frequently) sensitive-crime evidence. The canonical source for legal posture is Legal & Ethics SOP — never re-derive it here.

  • Platform ToS: bulk scraping, self-bots (user-token automation), mass DMs, coordinated inauthentic behavior, automated account creation, and ToS-circumventing client mods are prohibited by Discord's ToS and Community Guidelines. Investigation accounts that violate ToS risk ban and evidence-admissibility challenges.
  • Sock-puppet & pretexting framing: lurking in a server joined via a public invite is generally OSINT; sustained false-identity engagement (fake vendor persona, fake buyer interaction, paid VIP-tier access under a manufactured backstory) crosses into pretexting under several legal frameworks (US — varies by state and engagement context; EU — GDPR Recital 47 / lawful basis; UK — CMA pretexting prohibitions) and may compromise evidence admissibility. Document the legal basis in Collection Log.
  • DiscordChatExporter ToS gray area: self-owned data is fine; authorized customer-server engagement is fine; running it against a third-party server the analyst joined as a lurker is a ToS gray area — flag and route to Legal & Ethics SOP. Prefer authorized-bot collection when in scope; document the authority chain.
  • Wiretap / eavesdropping statutes: surreptitious recording of voice channels and Stage Channels triggers wiretap statutes in many jurisdictions (US — federal Wiretap Act and state-level "two-party consent" rules in CA, FL, IL, MA, MD, MT, NH, PA, WA; EU — varies); only record under explicit consent or legal process.
  • Webhook & token interaction limits: capturing a leaked webhook URL is OSINT; POSTing to it is unauthorized access. Capturing a leaked bot token is OSINT; using it to authenticate is unauthorized access. Treat any leaked credential as evidence to disclose, not to validate.
  • Sensitive-crime hard stop: CSAM, trafficking, and threat-to-life trigger immediate routing per Sensitive Crime Escalation SOP. Do not download; preserve URL + timestamp + snowflake-ID metadata only.
  • OPSEC: investigation phone numbers, payment methods, devices, and IPs must be isolated from the analyst's personal identity per OPSEC Planning SOP. Discord's coordinated-inauthentic-behavior detection links accounts via shared payment, shared IP, and device fingerprint; account warming + payment isolation + connection routing are required for sustained investigation accounts.
  • Connections-panel disclosure: capturing a target's voluntarily-public Connections (Twitch / GitHub / Spotify / etc.) is OSINT; correlating connections to deanonymize a pseudonymous target may engage GDPR/CCPA personal-data scope — ensure lawful basis is documented.

18) External Resources​

Official Documentation:

Third-Party Tools:

Server Discovery Directories:

Training & Guides:

Last Updated: 2026-04-27 Version: 1.0 (initial draft — Group H queue) Review Frequency: Quarterly (Fast tier — platform APIs, ToS, attacker TTPs rotate frequently)