FAQ

Common questions about StreamLinkr - setup, AI, keys, and platforms.

Back to Setup Guide

Getting Started

5
No - Streamer.bot is required. It handles the Twitch event feed (chat messages, raids, subs, channel points, etc.) and executes in-chat actions. StreamLinkr connects to it over WebSocket to receive those events and send replies back. The two work as a pair.
Through Streamer.bot. Instead of asking you to hand over your platform credentials to another third-party app, StreamLinkr routes everything through the automation tool you already use. If Streamer.bot can see the chat, StreamLinkr can read it and respond to it. Your platform logins stay inside Streamer.bot where you already trust them.
The StreamLinkr desktop app and Streamer.bot are both Windows-only. If you use Ollama as your AI provider, Ollama itself can run on Mac or Linux on a separate machine - StreamLinkr just needs to reach it over the network.
Nightbot, StreamElements, and Fossabot are command-and-response bots - they react to specific triggers with preset text. StreamLinkr is an AI layer that reads conversation context, remembers viewers, and generates replies based on what's actually happening in chat. It requires Streamer.bot and your own AI provider key (Ollama, Gemini, OpenAI, or Grok). Trade-off: more setup, more personality control, €10 once instead of a monthly subscription.
Download the latest installer from the Setup page and run it over your existing installation. The installer preserves your config.json and all database files. No manual migration needed - any new config keys are auto-synced from the example file on first launch.

Pricing & Keys

6
StreamLinkr is BYOK - Bring Your Own Key. The €10 covers access to StreamLinkr itself. For AI, it connects directly to your own API accounts (OpenAI, Gemini, Grok) or your local Ollama setup. You pay the AI providers only for what your chat actually consumes - typically pennies per stream for normal use. There is no hidden subscription or token markup on our end.
License keys are distributed via Ko-Fi. Purchase a key from the Ko-Fi shop, then sign in here with Google using the same email - your key will appear in the Dashboard automatically. If it isn't showing up, open a ticket on Discord.
Three steps:
  1. Buy a key from the Ko-Fi shop using an email address you control.
  2. Sign in here with Google using that same email - Ko-Fi has no login of its own, so Google is the bridge.
  3. Your key appears automatically in the Dashboard within seconds of signing in.

If your key isn't showing, make sure the Google account email matches the email you used on Ko-Fi. If they don't match, open a ticket on Discord.

Yes. Sign in with Google on the Dashboard and click Start a free trial - a trial key is generated for you instantly, no purchase or card required. Trial keys have the same features as a paid license and run for 7 days. Only one trial is allowed per account, and the option may be temporarily unavailable if trial key generation has been disabled.
Trial keys have the same features as a paid license, but two things are trial-only: the bot occasionally posts a short StreamLinkr promo message in chat (roughly every 45 minutes, only while the stream is live) as a condition of the free trial, and validation sends a locally SHA-256-hashed device fingerprint - required for trial keys, optional for paid ones - so a single trial can't be claimed on one machine and reused on another. No raw hardware identifiers ever leave your PC. Both go away the moment you upgrade to a Standard key.
The bot stops validating once the trial's expiry date passes - the Dashboard will show it as expired. To keep using StreamLinkr, buy a key on Ko-Fi using the same email as your Google account; the license appears in your Dashboard automatically alongside the expired trial. Trials can't be renewed or reclaimed once used.

AI & Performance

3
Ollama (local, free - runs entirely on your hardware), Google Gemini, OpenAI (GPT-4o, GPT-4o-mini, etc.), and Grok (xAI). You only need one. Ollama requires no API key and has no per-request cost - the only cost is electricity.
No. Ollama is completely free and runs locally - recommended for most streamers. If you prefer a cloud provider, Gemini has a generous free tier. OpenAI and Grok require a paid API key. Whichever you choose, you can switch anytime without restarting the bot.
No. Because StreamLinkr integrates with Streamer.bot and offloads the AI work to cloud API keys or a separate Ollama instance, the desktop app itself consumes almost no local resources - it's handling text requests and webhooks, not doing heavy computation. Your game and encoder performance are untouched. The only exception is running a local Ollama model on the same machine as your stream - in that case, you'll want a capable GPU so the LLM does not compete with your game for VRAM.

Platforms

3
StreamLinkr supports Twitch, Kick, and YouTube. You can configure one platform or more than one, depending on where you stream. Some integrations remain platform-specific where the underlying platform feature is unique.
Yes. StreamLinkr is designed around per-platform configuration, so you can enable multiple platforms and control command access separately for each one.
Command permissions are configured per platform. Check the command permission matrix and confirm the relevant Twitch, Kick, or YouTube roles are allowed for that command.

Features

6
Yes. You can edit the bot identity, tone, speech style, guidelines, response filters, and persona presets from the app. Those changes are live-configured, so you can refine the bot without editing code.
lurk posts a lurk message for the viewer, unlurk welcomes them back, and stfu temporarily stops the bot from replying for a specified number of minutes - commands continue to work during that window.
StreamLinkr can generate shoutouts automatically from events such as raids and subscriptions. When available, it uses profile and recent-stream context so the result is more personal than a static template.
Yes. With Speaker.bot configured, StreamLinkr can send replies and selected events to TTS. You can use !tts, !comment, random TTS triggers, and configured event announcements.
No. Settings are saved into config.json and hot-reloaded automatically by the app. In normal use, change the setting, save it, and the new value becomes active without restarting.
Use the database backup tools in StreamLinkr. The app supports automated backup behavior as well as manual backup and restore flows for persisted data such as memory and profiles.

Troubleshooting

8
Work through this checklist in order:
  1. Streamer.bot WebSocket server status shows Running.
  2. Address, Port, Endpoint, and Password in StreamLinkr match Streamer.bot exactly.
  3. The chatbot is started - status indicator in StreamLinkr should be green.
  4. Your bot Twitch account is connected in Streamer.bot under Platforms → Twitch.
  5. The Action Group was imported into Streamer.bot (see Connections).
No. A process lock file prevents duplicate instances from starting. This is intentional - running two instances against the same Twitch channel would produce duplicate bot messages in chat and corrupt the shared database state.
Greeting state is per-session - each viewer is greeted once per stream. The flags reset automatically when a StreamOnline event fires. If the bot is restarted mid-stream without a StreamOnline event, the flags reset on restart too, which can cause a second greeting. This is expected behaviour.
The anti-repetition system tracks recent responses per request type in the database and injects a "don't repeat" instruction into the prompt automatically. If responses are still repeating, check that:
  • The database is enabled in Settings → Database.
  • The database hasn't been recently purged.
  • Your AI model has enough context to follow the instruction (very small models may ignore it).
Add the bot account names to the Bot Ignore List in Settings. Common bots - Nightbot, StreamElements, StreamLabs, Fossabot, Moobot - are pre-populated by default. Any account name on the list is silently ignored regardless of what it says.
Start with the Logs page in the desktop app. It is usually the fastest place to spot connection errors, provider failures, permission issues, or runtime problems.
The backend uses a request queue to handle concurrent events. Requests are processed in order rather than all at once, which prevents duplicate replies, avoids hitting chat API rate limits, and keeps responses clean even during hype trains or raid bursts.
  1. Confirm the Twitch streamer account (not bot account) is connected in Streamer.bot under Platforms → Twitch.
  2. In Streamer.bot, go to Platforms → Twitch → Channel Point Rewards and verify the reward appears.
  3. Right-click the reward and select Copy Reward ID.
  4. Paste that ID into the correct field in StreamLinkr's Permissions tab for the action you want to trigger.
Still stuck?

Open a ticket on Discord - we'll sort it out.

Join Discord

We use Google Analytics to understand how StreamLinkr is used. This sets cookies only if you accept - see our Privacy Policy for details.