Bouncer

Appearance

🥇 Google AI Studio · Gemini 2.5 Flash (free) ​

Recommended #2. Gemini 2.5 Flash has a huge context window and strong reasoning. It's a great fit for the 🧠 Executive Summary in rescue handovers.

1. Get a free key ​

  1. Open aistudio.google.com/apikey and sign in with a Google account.
  2. Click Create API key and copy it (it starts with AIza).

2. Connect it (one line) ​

bash
bouncer config --preset gemini-gold --key AIza_your_key_here

Or keep the key in your environment (GEMINI_API_KEY or GOOGLE_API_KEY):

bash
export GEMINI_API_KEY=AIza_your_key_here
bouncer config --preset gemini-gold

Preset details ​

Endpointhttps://generativelanguage.googleapis.com/v1beta/openai/chat/completions
Modelgemini-2.5-flash
Level🥇 3, confidence gate 0.80
Sent per promptprompt text + up to 40 file paths (never file contents)

Privacy on the free tier

Google's terms for unpaid Gemini API usage allow prompts to be used to improve Google products. If your prompts are sensitive, use a paid key or Ollama instead.

Latency

Flash usually answers within Bouncer's 1500 ms budget. A slow reply simply falls back to Level 1 for that prompt. Run bouncer stats to see how often that happens ("fail-opens"), and raise the budget with bouncer config --timeout 2500 if needed.

Troubleshooting ​

  • failed (HTTP 400/403): the key is invalid, or the Gemini API isn't enabled for its project. Create a fresh key in AI Studio.
  • failed (HTTP 429): free-tier quota reached. Bouncer stays on Level 1 until it resets.
  • Remove it: bouncer config --clear.
Was this page helpful?