UNPKG

openclaw

Version:

Multi-channel AI gateway with extensible messaging integrations

71 lines (53 loc) • 3.28 kB
--- summary: "How OpenClaw loads environment variables and why service starts lose them" title: "Env vars and .env loading" read_when: - You are setting API keys through env or .env - Your env vars disappeared after starting the Gateway as a service --- ## Env vars and .env loading <AccordionGroup> <Accordion title="How does OpenClaw load environment variables?"> OpenClaw reads env vars from the parent process (shell, launchd/systemd, CI, etc.) and additionally loads: - `.env` from the current working directory. - a global fallback `.env` from `~/.openclaw/.env` (`$OPENCLAW_STATE_DIR/.env`). Normally, neither `.env` file overrides existing env vars. For an OpenClaw-installed systemd service, the global `.env` may replace only service values that OpenClaw recorded as managed; operator-owned service values still take precedence. Provider credential and endpoint-routing keys are an exception for workspace `.env`: keys such as `GEMINI_API_KEY`, `XAI_API_KEY`, `MISTRAL_API_KEY`, or any key ending in `_ENDPOINT` (and other bundled-provider auth or endpoint env vars) are ignored from workspace `.env` and should live in the process environment, `~/.openclaw/.env`, or config `env.vars`. Inline env vars in config apply only if missing from the process env: ```json5 { env: { vars: { OPENROUTER_API_KEY: "sk-or-...", GROQ_API_KEY: "gsk-...", }, }, } ``` See [/environment](/help/environment) for full precedence and sources. </Accordion> <Accordion title="I started the Gateway via the service and my env vars disappeared. What now?"> Two fixes: 1. Put the missing keys in `~/.openclaw/.env` so they load even when the service does not inherit your shell env. 2. Enable shell import (opt-in convenience): ```json5 { env: { shellEnv: { enabled: true, timeoutMs: 15000, }, }, } ``` This runs your login shell and imports only missing expected keys (never overrides). Env var equivalents: `OPENCLAW_LOAD_SHELL_ENV=1`, `OPENCLAW_SHELL_ENV_TIMEOUT_MS=15000`. </Accordion> <Accordion title='I set COPILOT_GITHUB_TOKEN, but models status shows "Shell env: off." Why?'> `openclaw models status` reports whether **shell env import** is enabled. "Shell env: off" does **not** mean your env vars are missing - it just means OpenClaw will not load your login shell automatically. If the Gateway runs as a service (launchd/systemd), it will not inherit your shell environment. Fix by putting the token in `~/.openclaw/.env`, enabling `env.shellEnv.enabled: true`, or adding it to config `env` (applies only if missing), then restarting the gateway and rechecking: ```bash openclaw models status ``` Copilot activates only with an explicit `models.providers.github-copilot` entry, a saved Copilot auth profile, or `COPILOT_GITHUB_TOKEN`. Generic `GH_TOKEN` and `GITHUB_TOKEN` variables do not enable or authenticate Copilot. Run `openclaw models auth login --provider github-copilot` to sign in. See [/concepts/model-providers](/concepts/model-providers) and [/environment](/help/environment). </Accordion> </AccordionGroup>