Skip to content

UI and Configuration Options

Jump to bottom
alex edited this page May 23, 2026 · 35 revisions

UI and Configuration Options

AI Tab

AI tab This is the main/default tab.

  • Start / Stop Services: Toggles the AI stack on/off. Stop and restart here after any service setting change. Services do not reload automatically.
  • Calibrate: Calibrates the app to your room noise level, microphone, and audio setup. Run on first use and after any hardware change.
  • OBS Overlay: When ON, the app displays an animated black window showing text from your input and AI responses. Add it to OBS as a screen capture, apply a Chroma Key filter, set the color to black, and set Similarity, Smoothness, and Key Color Spill Reduction to 0. Note: if the window is relaunched, it must be re-added in OBS.
  • Listen to me / Ignore me: Toggles input monitoring. When ignoring, the app does not respond unless trigger words such as "ship" or "listen" are detected. Say "stop ignoring me" to resume.
  • Detailed log on/off: Toggles detailed logging.

Player Tab

Player

  • Commander Name: If your in-game name is difficult for Text-to-Speech to pronounce, enter an easier alternative here.
  • Journal and Bindings Directory: On Windows this defaults to the standard location automatically. On Linux, set up symlinks as described in the installation guide.
  • Fleet Management: Assign voices, personalities, and cadence to individual ships. Personality and cadence require a cloud LLM. Settings can also be changed via voice command: "Change Name to George."

Settings / Local LLM Tab

Settings - Local LLM

  • Set the address of your inference server. Defaults to localhost with the Ollama URL.
  • Provide the names of the models to use. See the Local LLM guide.
  • LLM host radio buttons: Select between Ollama and LM Studio.
  • Use checkbox: Enable to use the local model instead of the cloud.

Settings / Audio

Settings - audio

  • Speech Volume: Controls the volume of speech synthesis.
  • TTS Voice Speed: Controls the speed of speech synthesis.
  • Beep Volume: Controls the volume of the beep indicator. Indicates that STT has finished processing and the LLM has received input.
  • STT Threads: Sets the thread allocation for STT processing. This is a min/max setting. The app requests the minimum but uses what the processor provides. Threads are released after processing completes.
  • Use Local Text To Speech: Overrides the cloud TTS key and uses local TTS.
  • Audio Wave Visualizer: Displays a dynamic graph of the audio input. Shows the noise floor, audio signal, gate zones, and clipping if present.

Settings / Cloud LLM Tab

Settings - Cloud LLM

  • Cloud LLM Key: Enter your API key. Supported providers: Gemini, OpenAI, Grok, Mistral, Deepseek, and Anthropic/Claude.
  • Cloud TTS Key: Enter your API key. Supported provider: Google.
  • Note: Uncheck the "Use" checkbox in Local LLM. It overrides the cloud LLM key.

LLM (AI Brain)

Cloud option: Enter your API key for Mistral, xAI, OpenAI, or Anthropic/Claude. The app uses a fixed model per provider:

  • Mistral: 'mistral-small-2506' (Free with hourly limit)
  • xAI: grok-4-1-fast-non-reasoning
  • OpenAI: gpt-4.1-mini (commands) / gpt-5.2 (queries)
  • Gemini Generative Language API: gemini-3.1-flash-lite-preview for commands and queries
  • Anthropic/Claude

Local option: Leave the key blank, fill in the local LLM fields below, and check ☑ Use next to the local LLM. See Local LLM guide (Linux) / Local LLM guide (Windows).

  • LLM Address: defaults to localhost. Replace with the IP of another PC if Ollama runs on a separate machine.
  • Command LLM: handles voice command interpretation.
  • Query LLM: handles data analysis. tulu3:8b is the minimum. Larger models produce better results.

No local hardware? Use a cloud LLM.

Cost will be different depending on which Cloud service you choose, and how long you play.

FREE CLOUD Option: Mistral

  1. Go to Mistral Console
  2. Create account with a valid email you can verify.
  3. NO CREDIT CARD NECESSARY
  4. Create a "Organisation" (Call it whatever you want. For example "Elite Intel")
  5. Generate an API key. Enter that key in to the app and restart the app.

Option A: xAI API Key

  1. Go to the xAI Console.
  2. Sign up or log in.
  3. Navigate to the API section and generate a new API key.
  4. Add credits to your account.
  5. Paste the key into the LLM field and check the lock box.

Option B: OpenAI API Key

  1. Go to the OpenAI Platform.
  2. Sign up or log in.
  3. Navigate to the API section and generate a new API key.
  4. Paste the key into the LLM field and check the lock box.

Option C: Anthropic/Claude API Key

  1. Go to the Claude Platform.
  2. Sign in with email or Google. Note: authentication uses a magic link sent to your email.
  3. Go to Settings → Billing and add credits before creating a key. A key created on an unfunded account does not function even if credits are added afterward.
  4. Go to API Keys and create a key.
  5. Paste it into the LLM field, check the lock box, and start or restart services on the AI tab.

Getting a Google TTS Key (14 voices)

  1. Go to the Google Cloud Console.
  2. Sign in or create an account.
  3. Create a new project.
  4. Enable the Generative Language API for LLM and/or Cloud Text-to-Speech API for TTS.
  5. Go to Credentials, create an API key, and copy it.
  6. Restrict the key: Click the key you just created. On the key detail page, click Restrict key. A dropdown appears. Check each API you enabled (STT and/or TTS), then click Save.
  7. Paste the key into the Speech to Text and/or Text to Speech fields in the app. Check the lock boxes.

App Settings and Data Directory

App settings and data are stored in a SQLite database located at:

  • Linux: ~/.local/share/elite-intel/elite-intel/db/
  • Windows: %APPDATA%\elite-intel\db\

For issues, contact via Matrix. Bug reports and pull requests are welcome.

Community 👉Matrix👈

Clone this wiki locally