TeleportGet hired faster

Use your own AI key

Connect an OpenAI-compatible or Anthropic API key for deeper fit analysis, full resume rewrites and open-ended answers.

Teleport works without AI. Adding your own key gives deeper fit analysis, full resume rewrites, written insight summaries and open-ended answers in Ask a question.

The AI engine settings with preset, API style, base URL, model and key
Pick a preset or enter any OpenAI-compatible endpoint.

Set it up#

  1. Open the dashboard and go to AI engine.
  2. Choose a Preset, or fill in API style, Base URL and Model yourself.
  3. Paste your API key (not needed for a local model).
  4. Check Use my AI model for fit analysis.
  5. Click Save. Chrome asks for permission to reach that one API address. Allow it.
  6. Click Test connection to confirm it works.

Presets#

PresetAPI styleDefault model
OpenAIOpenAI-compatiblegpt-4o-mini
AnthropicAnthropicclaude-3-5-haiku-latest
xAIOpenAI-compatiblegrok-3-mini
OpenRouterOpenAI-compatibleopenai/gpt-4o-mini
Local (Ollama)OpenAI-compatiblellama3.1, at http://localhost:11434/v1

Any service with an OpenAI-compatible chat completions API should work. Set its Base URL and Model.

Your key stays on your device#

  • The key is stored in your browser's local extension storage.
  • It's never sent to Teleport and never synced, and it's redacted when you export your data.
  • Requests go directly from your browser to your provider, under that provider's terms. Usage may be billed to you by the provider.
  • Forget key deletes it.

If something goes wrong#

If an AI request fails, Teleport falls back to the local engine and shows the error. Common causes:

  • The key is wrong or has no credit. Test connection shows the provider's error.
  • The permission prompt was declined. You'll see "Permission to reach the AI API was declined, so the local engine will be used." Click Save again and allow it.
  • A local model isn't running. Start Ollama and try again.