managed open-model instances

Deploy open
LLMs on a
managed instance.

One OpenAI-compatible endpoint, 42+ open models, and flat monthly pricing — no per-token bills. Deploy a model, get an API key, and call it from the tools you already use.

Free tier · 250K tokens/mo · nothing to manage

deepseek logo
qwen logo
openai logo
kimi logo
// the platform

The open-model API, without the per-token roulette.

One endpoint, a curated catalog of open models, and a bill you can predict. Built for developers who just want to ship.

One OpenAI-compatible endpoint

A single /v1/chat/completions endpoint that works with the OpenAI SDK and any tool you already use. Change the base URL and key — that's the whole migration.

42+ open models

DeepSeek, Qwen, Llama, GLM, Kimi, MiniMax, gpt-oss, Gemma and more — deploy any of them onto your instance. Each is routed to the cheapest provider serving it.

Flat monthly pricing

One monthly fee per instance tier — no per-token billing, no metered surprises. You always know the bill before the month starts.

Per-model keys & honest limits

Every deployed model gets its own API key. A dashboard shows your live throughput and monthly capacity, and your tier's limits are stated plainly — no hidden throttling.

// the catalog

42 open models, curated and ready.

The models actually worth running — each routed to the cheapest provider serving it. Deploy any on your instance.

deepseek logo

DeepSeek

· 7
DeepSeek V4 ProDeepSeek V4 FlashDeepSeek V3.2DeepSeek V3.1DeepSeek V3.1 TerminusDeepSeek V3 0324DeepSeek R1 0528
qwen logo

Qwen

· 9
Qwen3.7 MaxQwen3.5 397B A17BQwen3.5 122B A10BQwen3 Coder 480BQwen3 Next 80BQwen3 235B Instruct 2507Qwen3 235B A22BQwen3 32BQwen2.5 72B

Meta

· 4
Llama 4 MaverickLlama 4 ScoutLlama 3.3 70BLlama 3.1 8B

Google

· 3
Gemma 4 31BGemma 4 26B A4BGemma 3 27B

Mistral

· 2
Mistral Small 3.2Mistral Nemo
openai logo

OpenAI

· 2
gpt-oss 120Bgpt-oss 20B

Z.ai

· 6
GLM 5.2GLM 5.1GLM 5GLM 4.7GLM 4.7 FlashGLM 4.5 Air
kimi logo

Moonshot

· 3
Kimi K2.7 CodeKimi K2.6Kimi K2.5

MiniMax

· 4
MiniMax M3MiniMax M2.7MiniMax M2.5MiniMax M1

Nvidia

· 1
Nemotron 3 Nano 30B

Microsoft

· 1
Phi-4
// how it works

From model to endpoint in three steps.

01

Pick a model

Choose from 42 open models in one catalog — no juggling accounts and keys across five different providers.

02

Deploy it

Deploy onto your instance and get a scoped API key in seconds. No Dockerfiles, no GPUs, no DevOps to manage.

03

Call your endpoint

Point your OpenAI SDK at the endpoint with your key. Flat monthly cost — no surprise token bill at the end of the month.

ghosterr — your endpoint

$ curl https://gateway.ghosterr.io/v1/chat/completions \

  -H "Authorization: Bearer sk-ghosterr-…" \

  -d '{"model":"deepseek-v3.2","messages":[…]}'

◆ same schema as the OpenAI API

✓ { "choices": [{ "message": { "content": "Hello!" } }] }

  billing  flat monthly · no per-token charges

→ swap the base URL & key, keep your code

42
open models, one endpoint
OpenAI
compatible · drop-in API
$0
free tier — no card
Flat
monthly — no token bills
// pricing

Flat monthly pricing. Start free.

Pick an instance tier — each is a flat monthly fee with a monthly token allocation and a throughput ceiling. No per-token billing, ever. Change tiers any time.

Small Instance

$29/mo

Runs the essential open-source models — great for light and testing workloads.

Monthly capacity: Standard capacity

10 modelsyou can deploy · 7 families

QwenMetaGoogleMistralOpenAINvidiaMicrosoft

Medium Instance

$79/mo

More monthly capacity, with larger models unlocked including DeepSeek V3.

Monthly capacity: High capacity

23 modelsyou can deploy · 9 families

DeepSeekQwenMetaGoogleMistralOpenAIZ.aiNvidiaMicrosoft

Large Instance

$199/mo

Top capacity with every supported model unlocked, including the largest.

Monthly capacity: Max capacity

42 modelsyou can deploy · 11 families

DeepSeekQwenMetaGoogleMistralOpenAIZ.aiMoonshotMiniMaxNvidiaMicrosoft
custom

Enterprise Instance

Custom/ let's talk

Dedicated capacity, higher throughput, priority support, and custom model access — tailored to your workload.

Monthly capacity: Negotiated

Everything in Large,plus:

  • Custom throughput & monthly capacity
  • Dedicated support & SLAs
  • Volume pricing & custom model access
Contact sales →
Every paid tier starts from a permanent Free plan — no card required to sign up.

Deploy your first model free.

Sign up, deploy an open model, and call it from your code in minutes. No card, no infra to manage — the Free tier is permanent.

42+ open models · OpenAI-compatible · flat monthly pricing