The smallest connectable eve agent: a standard authored agent folder you can build and talk to locally.
hello-eve/
├── .env # ANTHROPIC_API_KEY (a DeepSeek key — see agent/agent.ts)
├── package.json
├── tsconfig.json
└── agent/ # eve's nested authored layout
├── agent.ts # model config (DeepSeek via the Anthropic-compatible API)
├── instructions.md # the always-on system prompt
├── sandbox.ts # just-bash execution backend
└── channels/
└── eve.ts # HTTP/NDJSON channel, no auth (local dev only)
agent/agent.ts points the Anthropic client at DeepSeek's Anthropic-compatible
endpoint (https://api.deepseek.com/anthropic), which reads ANTHROPIC_API_KEY.
So the key in .env is a DeepSeek key, and that is what lets the agent make
outbound model calls. Swap the baseURL/model for real Anthropic if you prefer.
pnpm install
pnpm dev # eve dev — starts the local agent server + TUIThen talk to it over the eve channel (default dev server on port 3000):
curl -X POST http://localhost:3000/eve/v1/session \
-H 'content-type: application/json' \
-d '{"message":"Reply with exactly: pong"}'This project uses eve's built-in local world, so no database is required —
state is in-process and resets when the agent restarts. To make workflows
durable, add experimental.workflow.world: "@workflow/world-postgres" to
agent/agent.ts, add the @workflow/world-postgres dependency, and set
WORKFLOW_POSTGRES_URL to a reachable Postgres.