Limits and quotas

Every per-surface quota and rate limit in one place — Agent sizes and provisioning time, Plain VM tiers and firewall rule counts, glamp tarball size, virtual key minimum spend cap, OAuth token TTLs, and refresh timeouts. Numbers are stated with units and refreshed on every deploy.

Last updated: 2026-06-09

Every number on this page is stated in concrete units. Hedged adjectives like "fast" or "affordable" are not allowed — see the limit and decide for yourself.

Agents

Limit	Value
Sizes	small, medium, large
Regions	US, EU
Provisioning time	~60 seconds
Agent name	non-empty; sanitized into the URL subdomain (≤30 chars)

Plain VMs

Limit	Value
Tiers	small, medium, large
Regions	US, EU
OS images	Ubuntu 24.04 LTS, Ubuntu 22.04 LTS, Debian 13, Debian 12, Fedora 43, Rocky 10, Alma 10
Firewall rules per VM	50
Provisioning time	~60 seconds
Stopped time	bills the same as running (Hetzner charges either way)

App VMs

Limit	Value
Sizes	small, medium, large
Regions	US, EU
App name	1–63 chars, lowercase alphanumeric and hyphen
Container port	1–65535
Environment variable value	up to 32768 chars each
Persistent volume	optional, mounts at /var/lib/app/data, lost on destroy
Stopped time	bills the same as running

GlampHub

Limit	Value
Max compressed tarball size	500 MB
Glamp name	kebab-case, 1–50 characters

LLM proxy and voice

Limit	Value
Supported providers	OpenAI, Anthropic, ElevenLabs
Minimum spend cap	$20
Usage reporting cadence	within ~2 minutes
Virtual key prefix	cg-pk-

App integrations

Limit	Value
Supported providers	Google, GitHub, Spotify
Refresh fetch timeout	10 seconds
Authorization window	10 minutes

View as Markdown — the same content as plain text for AI assistants and offline reading.

Was this helpful?