cube-js/cube · error

context `{name}` not found in config (run `cube login --cont

Error message

context `{name}` not found in config (run `cube login --context {name}`)

What it means

The CLI resolves authentication from a named context in its config file. When the user explicitly passes --context <name> (or CUBE_CONTEXT) but no context with that name exists in the config, `api()` bails with this message suggesting `cube login --context <name>` to create it.

Source

Thrown at rust/cube-cli/src/main.rs:70

}

impl Ctx {
    fn new(global: &GlobalArgs) -> Result<Self> {
        Ok(Self {
            json: global.json,
            config: config::Config::load()?,
            api_url: global.api_url.clone(),
            token: global.token.clone(),
            context: global.context.clone(),
        })
    }

    /// Build an authenticated API client from flags, env, or the config file.
    pub fn api(&self) -> Result<client::Client> {
        let ctx = self.config.context(self.context.as_deref());
        if let Some(name) = &self.context {
            if ctx.is_none() {
                bail!("context `{name}` not found in config (run `cube login --context {name}`)");
            }
        }
        let url = self
            .api_url
            .clone()
            .or_else(|| ctx.map(|(_, c)| c.url.clone()))
            .map(|u| util::normalize_url(&u));
        // An explicit --token / CUBE_API_KEY wins and disables auto-refresh
        // (it isn't tied to a stored refresh token).
        let token = self
            .token
            .clone()
            .or_else(|| ctx.map(|(_, c)| c.api_key.clone()));
        match (url, token) {
            (Some(url), Some(token)) => {
                // Enable auto-refresh only when using the context's own access
                // token and it has a refresh token saved alongside it.
                let refresh = if self.token.is_none() {

View on GitHub (pinned to 7d981676b3)

Solutions

1. Run `cube login --context <name>` to create and authenticate the context
2. Check the exact context name in the CLI config file (~/.cube/config or equivalent)
3. Fix the --context flag value or the CUBE_CONTEXT env var typo

Example fix

// before
cube deployments list --context prodction
// after
cube login --context production
cube deployments list --context production

const ctx = process.env.CUBE_CONTEXT || flags.context;
const cfg = JSON.parse(fs.readFileSync(configPath, "utf8"));
if (ctx && !(ctx in (cfg.contexts ?? {}))) {
  throw new Error(`context '${ctx}' missing; run: cube login --context ${ctx}`);
}

const hasContext = (c: unknown): c is { contexts: Record<string, unknown> } =>
  typeof c === "object" && c !== null && "contexts" in c;

try {
  run(`cube deployments list --context ${ctx}`);
} catch (e) {
  if (String(e).includes("not found in config")) {
    run(`cube login --context ${ctx}`);
    run(`cube deployments list --context ${ctx}`);
  } else throw e;
}

When it happens

Trigger: Running any command with --context production when `cube login --context production` was never run; typos in the context name; switching machines without copying the CLI config; CI environments missing the config file.

Common situations: Multiple environments (staging/prod) managed via contexts; fresh laptop setup; sharing scripts that hard-code a context name.

Related errors

• Provided token isn't for ${dotCubeCloud.url}
• context `{name}` not found (run `cube login --name {name}`)
• not logged in: run `cube login`, or set CUBE_API_URL and CUB
• device code expired before it was authorized; run `cube logi
• authorization was denied in the browser

AI-assisted analysis of cube-js/cube@7d981676b3 (2026-09-02). Data as JSON: /api/errors/45504664777d6742. Report an issue: GitHub.