Starter 4 vCPU / up to 32 GB RAM
$59/mo
- 4 shared vCPU
- Up to 32 GB ECC RAM (shared)
- 100M AI tokens/mo
- 250 GB NVMe disk
- Upload up to 3 Mbit/s
Server RentalPrivate containers.
Private containers.
Shared compute.
Dev Node containers and the private Model API. Every plan is a monthly subscription through Stripe checkout on this site. The details below are the listed plans, not independently tested infrastructure performance.
New rentals subscribe monthly through Stripe checkout on this site. Existing subscribers keep their current access; email [email protected] for support.
Product 02
Dev Node Listed plans
The listed Dev Node service provides an isolated Linux container on a 96-core / 1 TB ECC workstation, reached privately over Tailscale. vCPUs are shared and burstable; RAM is oversubscribed with a hard per-container limit and no swap.
- Included AI tokens + Hermes Agent: every Dev Node comes with Hermes Agent preinstalled and included GLM-5.3-Flash Model API tokens, with 100M to 400M tokens a month depending on the plan. Use them with Hermes or any OpenAI-compatible client.
- Backups included: On-site hourly snapshots, kept up to 14 days; restores on request, best-effort, and not a substitute for your own backups.
- Fast NVMe storage: your own disk quota, from 250 GB to 2 TB.
- Listed container access: root inside an isolated gVisor container, with no access to the host or other customers. Execution is CPU-only.
- Shared, burstable vCPUs and oversubscribed RAM: up to your plan's limit (hard per-container cap, no swap), not set aside for you.
- Private access over Tailscale only, no public ports.
- Shared upload with per-tier upload limits and a 250 GB/month egress cap.
- Built for compute: development, CI, benchmarking and batch jobs.
- Model API plans: see the listed API tiers.
Monthly plans
Listed configuration is not a guarantee of live capacity or order acceptance.
Standard 8 vCPU / up to 64 GB RAM
$119/mo
- 8 shared vCPU
- Up to 64 GB ECC RAM (shared)
- 150M AI tokens/mo
- 500 GB NVMe disk
- Upload up to 4 Mbit/s
Dev Slice 24 vCPU / up to 128 GB RAM
$249/mo
- 24 shared vCPU
- Up to 128 GB ECC RAM (shared)
- 250M AI tokens/mo
- 1 TB NVMe disk
- Upload up to 6 Mbit/s
Dev Node 24 vCPU / up to 256 GB RAM
$399/mo
- 24 shared vCPU
- Up to 256 GB ECC RAM (shared)
- 400M AI tokens/mo
- 2 TB NVMe disk
- Upload up to 6 Mbit/s
Listed prices are USD per month, plus applicable tax, and renew until canceled. One month is due at checkout. Prices exclude VAT; applicable tax is calculated at Stripe checkout.
Listed service limits. CPU-only containers on shared resources; no uptime SLA. Our workloads have priority. Shared resources may provide less than your plan’s vCPU or upload rate. Pro-rated credit applies for sustained preemption or extended outages, and for host memory shortages that block use for over 24 hours in a billing period; see the rental terms and applicable service limits below. Available worldwide, except sanctioned or restricted regions (Cuba, Iran, North Korea, Syria, Russia, Belarus and occupied regions of Ukraine); see Terms.
Listed included tokens: Hermes Agent preinstalled; 100M/month on Starter, 150M on Standard, 250M on Dev Slice and 400M on Dev Node. Starter and Standard allow 2 requests/min; Dev Slice and Dev Node allow 4 requests/min, all one request at a time. Allowances reset monthly with no rollover; requests are refused at the limit until reset.
Backups: On-site hourly snapshots, kept up to 14 days; restores on request, best-effort, and not a substitute for your own backups. Keep your own off-site copies of critical data. vCPUs are shared and burstable, so speed can vary at busy times; RAM is oversubscribed: each container has a hard limit at its plan size with no swap, but the RAM limits of the Dev Nodes we sell can add up to more than the memory available to Dev Nodes on the host, so when host memory runs short, processes may be stopped and your node restarted before it reaches the limit; in-memory data is lost, files on disk are kept.
Rental terms: no hosting, seedboxes or mining. Terms, AUP and Privacy Notice apply; review the rental terms and applicable service limits before purchase. Report abuse to [email protected]. Sold by SlowGlow AI LLC (NC).
Product 01
Private Model API Listed plans
The listed API uses GLM-5.3-Flash with OpenAI-compatible access over Tailscale and an API key; no public endpoint. Context and stream limits depend on the plan.
- OpenAI-compatible endpoint, so existing SDKs and tools work with a base-URL change.
- Private access over Tailscale.
Monthly plans
Listed configuration is not a guarantee of live capacity or order acceptance.
Solo
$25/mo
- 500M tokens
- 1 concurrent stream
- 128K context
- 4 RPM, burst 8
Pro
$79/mo
- 2B tokens
- 2 concurrent streams
- 256K context
- 15 RPM
Reserved 512K
$300/mo
- No fixed monthly token cap (fair use)
- 512K context
- Up to 1 stream; sold to one customer at a time
Reserved 1M
$500/mo
- No fixed monthly token cap (fair use)
- 1M context
- Up to 1 stream; sold to one customer at a time
Listed prices are USD per month, plus applicable tax, and renew until canceled. One month is due at checkout. Prices exclude VAT; applicable tax is calculated at Stripe checkout.
Fine print. No uptime SLA; owner and client workloads take precedence. Capacity is shared; requests may queue or receive 503 with Retry-After when busy, and per-key limits return 429. Available worldwide, except sanctioned or restricted regions (Cuba, Iran, North Korea, Syria, Russia, Belarus and occupied regions of Ukraine); see Terms.
The model is NVIDIA's FP4 checkpoint, licensed under the NVIDIA Open Model Agreement. Base GLM-5.3-Flash is MIT licensed, © 2026 Z.AI. SlowGlow AI is not affiliated with or endorsed by NVIDIA. Sold by SlowGlow AI LLC (NC).
Let's talk
Questions about the listed API or Dev Node plans? Get in touch.