Skip to main content

E2E testing

Authenticate both providers:

# Subscription (opens browser for OAuth flow):
cargo run -p agentkit-switchboard -- --config crates/agentkit-switchboard/e2e.toml auth login chatgpt_sub

# API key (stores raw key in system keychain):
cargo run -p agentkit-switchboard -- --config crates/agentkit-switchboard/e2e.toml auth add openai_payg sk-...

Start proxy:

cargo run -p agentkit-switchboard -- --config crates/agentkit-switchboard/e2e.toml

Scenario A: Subscription routing​

curl -i http://127.0.0.1:3812/openai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5.4-mini","messages":[{"role":"user","content":"hi"}],"stream":false}'

Expected: HTTP 200, x-switchboard-provider: chatgpt_sub, x-switchboard-billing: subscription.

Scenario B: Session affinity​

curl -i http://127.0.0.1:3812/openai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-Session-Id: sess-e2e-1" \
  -d '{"model":"gpt-5.4-mini","messages":[{"role":"user","content":"hi"}],"stream":false}'

Run twice. Assert x-switchboard-provider is the same both times.

Cleanup​

cargo run -p agentkit-switchboard -- --config crates/agentkit-switchboard/e2e.toml auth logout chatgpt_sub
cargo run -p agentkit-switchboard -- --config crates/agentkit-switchboard/e2e.toml auth logout openai_payg