OpenAI Integration

Save your OpenAI secret key to your OS Keychain:

agentsecrets secrets set OPENAI_API_KEY=sk-proj-12345...

This automatically saves the secret to your local OS Keychain and synchronizes the encrypted ciphertext to your cloud workspace.

Authorize OpenAI in your workspace allowlist:

agentsecrets workspace allowlist add api.openai.com

With Transparent HTTP Client Interception, you can use the official openai package directly:

import openai
from agentsecrets import init, credential

# 1. Enable interception once at startup
init()

# 2. Pass the zero-knowledge credential placeholder
client = openai.OpenAI(api_key=credential.OPENAI_API_KEY)

# 3. Call OpenAI naturally — raw keys never enter application RAM!
chat_completion = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {"role": "user", "content": "Explain zero-knowledge credential proxies in one sentence."}
    ],
)

print(chat_completion.choices[0].message.content)

Why this is secure:

• Prompt Injection Defense: Even if an LLM is tricked into dumping initialized variables or inspecting memory, client.api_key only ever contains "AS_SECRET_OPENAI_API_KEY".
• Automatic Injection: Real keys are injected directly into the outbound TLS stream by the local proxy.
• Stream Support: Native SDK streaming (stream=True) works seamlessly without changes.