-
Notifications
You must be signed in to change notification settings - Fork 4
Home
Let’s jump into the cockpit and talk about Elite Intel—what it does, what it doesn’t, and how to make it your trusty wingman.
Elite Intel is an AI assistant for Elite Dangerous, built to understand your voice and help you navigate the galaxy without memorizing a 50-page command list like some other tools (looking at you, Voice Attack). It listens to what you say and figures out your intent. That’s both awesome and tricky. Awesome because you can just talk naturally—no need to recite “engage hyperdrive protocol 47-B.” Tricky because you gotta be clear with your words to get the results you want. I’ll toss in some examples later to show you what I mean.
First up, how does Elite Intel hear you? That’s Speech-to-Text (STT). It grabs your voice through your mic, chews it up, and spits it out as text for the AI to process. Pro tip: a decent mic makes a huge difference. The app auto-detects your audio settings and calibrates for the best results, but even with top-tier gear, it’s not perfect. We’re using Google’s Speech-to-Text engine right now, and like any STT, it can stumble over words or phrases. If it keeps hearing “southwest” when you say “set voice to,” you can fix it in the dictionary file (check the app’s dictionary directory). Just add your correction in the format: "southwest"="set voice to". Easy.
Once your voice is text, the AI decides if you’re giving a command, asking a question, or just chatting about the void. Say “lower the landing gear,” and it’ll flip that switch for you. Clear as a crystal shard. But if you say something vague like “prepare for landing,” it’s a coin toss—maybe a 70% chance it drops the gear, or it might just say, “Preparing for landing,” and do nothing. Why? Because the AI’s got moods. Seriously. Sometimes it needs you to be direct. Say “approaching planet,” and it’ll likely throttle down to 25% speed. That’s just how it rolls.
Text-to-Speech (TTS) is how Elite Intel responds, using one of 14 Google-powered voices (more might come later). Short commands like “deploy hardpoints” or "weapons hot!" usually get a quick action with maybe a brief “Got it.” But ask something meaty, like “What’s the security status of our next jump?” and it’ll give you a full rundown—assuming you’ve got a system targeted in your nav panel. You can phrase it however feels natural: “Give me the lowdown on security for our next jump,” “How safe is the system we’re jumping to?” or even “Any pirates ahead?” That last one’s a bit vague, so it might not always trigger the security check.
Here’s how it works: the app pulls data from EDSM for your targeted system, feeds it to the AI, and lets it analyze. If EDSM shows high traffic and recent commander deaths, it might say, “High security risk—possible pirate activity.” But it’s not omniscient. It can’t actually know if pirates are lurking, since that data’s not out there (at least not legally).
The Large Language Model (LLM) is Elite Intel’s smarts—its ability to reason, decode your ramblings, and keep the convo flowing. We’re rocking xAI brain power, and it’s a beast. It doesn’t just mash keyboard buttons; it can tap into its knowledge base for insights. Ask it to analyze your ship’s loadout, and it’ll break down the strengths and weaknesses of your build. It’s not telling you to rush to an engineer and rebuild your ship—it’s just giving its two credits based on what it knows. Your playstyle? That’s on you.
You can also ask stuff like, “Analyze local signals” or “What’s on the scanners?” The AI digs into in-app session data (like auto-scans or FSS discoveries) and gives you a summary. Try “Are there resource sites in this system?” or “Any starports nearby?” It’ll tell you what it finds, but here’s the catch: not all info is available. If it’s on your in-game panels but not in journal files or EDSM, the AI might shrug and say, “No data available” or “Nothing of interest here.” That’s because Elite Intel plays nice with the game’s Terms of Service and doesn’t peek at your in-game memory state. You see it, but the AI might not.
The user interface? Lean and mean. No endless menus or sliders—just three fields for the API keys, a few buttons to start services, and that’s it. Everything else happens through voice. Talk to it, mess around, ask weird questions, have fun. It’s your Sidekick, not your autopilot. Don’t expect it to fly the ship for you. Keep critical controls—speed, direction, FTL, chaff, weapons—in your hands. Why? Because stuff can go wrong. Network lag, Google STT mishearing you, a command I haven’t coded yet, or just a random bug could trip things up. Elite Intel is your buddy, not the commander.
Final Tips Experiment. Play with it. See what it can do. But always stay in control. If the AI misfires or doesn’t get you, no big deal—just keep flying. Got questions or quirks? Hit me up, and let’s make this app even better. o7, commander.
👉Discord.👈