NousResearch/hermes-agent · error · Error
real stream never started (no LLM credit / backend?)
Error message
real stream never started (no LLM credit / backend?)
What it means
Error "real stream never started (no LLM credit / backend?)" thrown in NousResearch/hermes-agent.
Source
Thrown at apps/desktop/scripts/perf/scenarios/stream.mjs:182
const el = document.querySelector(${JSON.stringify(SELECTORS.composer)})
el && el.dispatchEvent(new KeyboardEvent('keydown', { key: 'Enter', code: 'Enter', bubbles: true, cancelable: true }))
})()`)
const deadline = Date.now() + Number(opts.timeoutMs ?? 60000)
let started = false
while (Date.now() < deadline) {
await sleep(50)
const n = await cdp.eval(`document.querySelectorAll(${JSON.stringify(SELECTORS.assistantMessage)}).length`)
if (n > baseCount) {
started = true
break
}
}
if (!started) {
throw new Error('real stream never started (no LLM credit / backend?)')
}
await cdp.eval('window.__MO__.arm()')
// Let it run to completion or timeout.
const runDeadline = Date.now() + Number(opts.runMs ?? 30000)
while (Date.now() < runDeadline) {
await sleep(250)
const busy = await cdp.eval(`!!document.querySelector('[data-status="running"], [data-busy="true"]')`)
if (!busy) {
break
}
}
} else {
// Synthetic path: drive $messages directly. No LLM, no credits.
await cdp.eval(
`window.__PERF_DRIVE__.stream({ chunk: ${JSON.stringify(chunk)}, intervalMs: ${intervalMs}, totalTokens: ${tokens}, flushMinMs: ${flushMinMs} })`View on GitHub (pinned to c896c09c42)
Solutions
- Confirm the backend has LLM credentials and quota, then retry.
- Run with the synthetic stream path instead of a real LLM stream.
- Check gateway logs for the stream start failure.
When it happens
Trigger: Thrown at apps/desktop/scripts/perf/scenarios/stream.mjs:182 when the library encounters an invalid state.
Common situations: See trigger scenarios.
AI-assisted analysis of NousResearch/hermes-agent@c896c09c42 (2026-08-14).
Data as JSON: /api/errors/b7dc87d63f4e0ac8.
Report an issue: GitHub.