> ## Documentation Index > Fetch the complete documentation index at: https://docs.getsignalhub.com/llms.txt > Use this file to discover all available pages before exploring further. # Quickstart > Start building your first tracker in minutes The core concept of SignalHub is the **tracker**. A tracker monitors a specific topic and has its own sources, prompt, and settings. ## Follow / Fork a tracker The easiest way to get started is to follow or fork a tracker from the community library. You can explore the community library on [search](https://getsignalhub.com/search) to discover trackers shared by others. You can view each tracker's prompt, sources, settings, and updates. On any tracker, you can: * Follow: subscribe to its updates without managing any settings. * Fork: create your own copy, then customize the prompt, sources, and settings. The copy is fully managed by you. ## Create a tracker ### Step 1: Create a tracker * After signing in, you will be redirected to your [dashboard](https://getsignalhub.com/dashboard) where you can see your tracker list. * If this is your first sign-in, the list is empty. Click "New Tracker" to create one. For inspiration, browse the community library in [search](https://getsignalhub.com/search). ### Step 2: Basic settings Basic settings * Give your tracker a name to indicate its topic, for example "White House Watch: Immigration and Technology". * You can enable or disable the tracker. If disabled, this tracker will be paused and you won't receive any updates from it. * You can share your tracker with the community. When public, anyone can find it on [search](https://getsignalhub.com/search) and view its prompt, sources, settings, and updates. But your private notification channels are never shared. ### Step 3: AI settings AI settings * **Filtering**: Toggle to enable or disable AI-powered filtering. When on, messages will be filtered by the LLM based on your rules. * **Filter Rules**: The most important setting — it tells the tracker what you care about. For example: ``` [Your role / profession] - AI news journalist (global coverage) [Keep] - Major global AI developments with clear industry impact - Significant technical breakthroughs from leading AI companies - Important AI product launches or major feature updates - Major funding rounds involving high-impact AI startups - Focus areas: - Large language models (LLMs) - Multimodal models - Speech technologies (ASR automatic speech recognition,TTS text to speech) - Image, video, and audio generation models - Robotics and embodied AI - AI agents - Semiconductors and accelerators - Data centers and AI infrastructure - Data labeling and training pipelines - Reinforcement learning - Training and inference infrastructure [Filter out] - Pure marketing or promotional content - Minor business partnerships or small funding rounds - Opinion or commentary articles,unless written by globally recognized and well-established AI leaders ``` * **Filter Strictness**: Adjust how strictly the AI applies your filter rules. Higher strictness means more aggressive filtering, which may result in fewer signals but higher relevance. * **Rewrite Headline**: You can choose whether the AI rewrites headlines. We recommend leaving this enabled, as it helps reduce clickbait and misleading titles. * **Translate Updates**: You can enable translation and choose a target language to receive updates in your preferred language. ### Step 4: Notification settings Notification settings * Choose how you want to receive updates. Available channels include Email, Telegram, Discord, Slack, Microsoft Teams, Google Chat, Ntfy, Pushover, WeCom, Lark, DingTalk, and X/Twitter. * By default, updates are emailed to your sign-in address once per day at 7:30 AM. * For the best experience, add a real-time IM (Instant Messaging) channel (e.g., Telegram, Discord or Slack). See the channel setup guides in the Notifications section of the docs. ### Step 5: Sources settings Sources settings Manage the sources you trust. We recommend configuring 5-20 sources per tracker. SignalHub supports most websites, including many dynamic and UGC platforms. If a source fails to crawl, please contact us so we can improve support.