One layer that organises everything in one place

Your app talks to a name. What happens behind it, which model, which provider, how much privacy, you configure here and change any time, without touching a line of code. Try it:

ALLSmartLLM

Alias big-fat-model

Target

resolves to ALLSmartLLM · Qwen 3.5 · 397B

Privacy

Not needed. The prompt never leaves our house anyway.

What your app sends

Please draft the reply to Julia Berger, the new CFO at Innovatec Trading in Salzburg. She wants to know the status of the open invoice 2026-0417. Payment was sent to account AT61 1904 3002 3457 3201. For the ERP import please use the API key sk-live-9f2ac41b8e.

What the target sees, ALLSmartLLM

● Unchanged, the prompt stays on ALLSmartLLM in Austria.

^* Your key is stored encrypted at rest (AES-256, master key in the OS keyring, never in the database, never in the repo). Decrypted only in the memory of the request handler that is currently using it. Never logged, never returned via any endpoint, revocable in the portal in one click.

Own inference

Open-weights models on hardware we own, in Upper Austria. No sub-processors, no third countries, no reselling of third-party APIs.

One name, any model

An alias bundles model, provider and policy. It points at our own models, or, with your own key, outside. Switching models means: change it here, nothing else.

Privacy on tap

Per alias in stages: names, addresses, IBANs, and API keys, tokens and credentials that end up in prompts by accident. Only what's allowed to leave, leaves.

OpenRouter direct vs. via ALLSmartLLM

Both are valid paths. The difference is invoicing, privacy, and who you email when something breaks.

OpenRouter direct

Privacyno layer, prompt goes out as-is
ComplianceUS provider, no DPA
Rate limitper OpenRouter key
Model swapchange client code
SupportDiscord + docs
Fallbacknone, 500 is the end
Model catalogue~500 models instantly

via ALLSmartLLM

PrivacyGLiNER detection + LLM check on strict
ComplianceGDPR, DPA with ALL Smart GmbH
Rate limitunified across all providers in your account
Model swapreroute alias, client unchanged
SupportOne point of contact, German or English
FallbackMulti-provider chain automatic
Model catalogue~15 curated + BYOK pass-through to OpenRouter/OpenAI

Inference is the raw material. This is the layer that makes it usable.

The smart part was never the model. It's what surrounds it: routing, policy, privacy, proof.

One API, your aliases

An alias is more than a model name, it bundles model, provider and privacy policy. Swap the model centrally; your application never changes.

big-fat-model ──► kimi-k2.6 kimi-k3

// alias updated · zero code changes · apps keep running

Privacy in stages

Off, pseudonymize, or go strict: an LLM on our own hardware double-checks for indirect identifiers before anything leaves the EU.

aus mask pseudonymize strict

Please reply to <<PERSON_1>> from <<ORG_1>> regarding invoice <<IBAN_1>>

Mapping lives in request memory only. Never stored, never logged.

Your data stays in Europe

Our own hardware in Austria. No sub-processors, no third countries. External frontier models are reachable, but only through the masking layer.

your app ──► ALLSmartLLM · AT ─▍ third country

Prompts terminate on hardware we own, or leave masked, by your policy.

Measured, not promised

Nightly benchmarks with published methodology. Detection recall for the privacy layer as a number, not an adjective.

Qwen 3.5 397B

54.8 t/s

Nemotron 3 Nano Omni

47.8 t/s

Nemotron Cascade 2

47.9 t/s

Decode rate, streaming, median of repeated nightly runs.

Fallback chain for dead routes

Up to four targets per alias. If the primary provider flakes (429, 5xx, timeout), the request walks to the next one automatically. Across ALLSmartLLM and OpenRouter Free if you want.

free ──► google:free ──► nvidia:free ──► deepseek:free

// automatic, live mid-stream, no retry code on your side

Bring your own provider keys, encrypted

Plug in your OpenAI, Anthropic or OpenRouter account. Keys land AES-256-GCM encrypted in a vault, master key from the OS keyring. Plaintext lives in request memory only. Never in logs, never in backups.

dein Key ──► Vault · AES-256-GCM ──► Provider-Call

AES-256-GCM · OS keyring · zero-log for plaintext

Guards, rules before the answer

Rule-based policy checks per alias. „log_only" writes violations to the audit log, „enforce" breaks the request. For compliance proof and hard limits: no customer data into ext. models, no card numbers in the response stream, no PII into training data.

aus log_only enforce

log_only to observe · enforce to block

Not one provider, all of them

One key with us, every major upstream underneath: our own ALLSmartLLM hardware, OpenRouter (~300 models), OpenAI, Anthropic, Groq, DeepSeek. Switching provider means re-pointing an alias.

ALLSmartLLM OpenRouter OpenAI Anthropic Groq DeepSeek

6 providers · 1 API · mix per alias

First response in under five minutes

If your software speaks OpenAI, it already speaks ALLSmartLLM. Change one URL.

step 1

Create your account

10M tokens free to test and set up. No card, no subscription.

free tokens10,000,000

step 2

Get your API key

One key, all models, all aliases. Usage-based, nothing else.

SMARTLLM_API_KEY••••••••••••

step 3

Make your first request

Call an alias, not a model version. Privacy is one parameter.

curl https://allsmartllm.com/v1/chat/completions \
  -H "Authorization: Bearer $KEY" \
  -d '{
    "model": "frontier",
    "privacy": {"mode": "pseudonymize"},
    "messages": [...]
  }'

Later, host it yourself: same API on your own hardware. When your box is running, you change one URL, done.

Curated open-weights models

Vetted, benchmarked, and kept current. Aliases follow our recommendation, or pin a version and stay.

Model	Alias	Can do	Context	License	Quality^*	Measured	Status
Qwen 3.5 397B 397B / 17B MoE (int4 AutoRound)	frontier	ToolsBildReasoning	256K	Apache 2.0	26	54.8 t/s	loaded
Nemotron 3 Nano Omni 33B / 3B MoE (Q4_K_M)	-	ToolsBildReasoning	256K	NVIDIA Open Model License	9	47.8 t/s	on-demand First start 50 s
Nemotron Cascade 2 32B dense (Q4_K_M)	-	ToolsReasoning	256K	NVIDIA Open Model License	12	47.9 t/s	on-demand First start 50 s
Phi 4 14B dense (Q4_K_M)	-	Text	16K	MIT	1	15.2 t/s	on-demand First start 8,3 s
Qwen 3.6 35B 35B / 3B MoE (FP8)	mid vision	ToolsBildReasoning	256K	Apache 2.0	26	78.7 t/s	loaded
Qwen 3.5 35B 35B / 3B MoE (Q4_K_M)	-	ToolsBildReasoning	256K	Apache 2.0	23	41.6 t/s	on-demand First start 54 s
Granite 4.1 30B dense (Q4_K_M)	-	Tools	256K	Apache 2.0	3	7.6 t/s	on-demand First start 34 s
Nemotron 3 Nano 30B / 3.5B MoE (Q4_K_M)	-	ToolsReasoning	256K	NVIDIA Open Model License	9	48.3 t/s	on-demand First start 9,0 s
Nemotron 3.5 Lightning 30B / 3B MoE (NVFP4 (ModelOpt, MIXED_PRECISION))	-	ToolsReasoning	256K	OpenMDW-1.1	16	132.4 t/s	loaded
Qwen 3.6 27B 27B dense (Q4_K_M)	-	ToolsBildReasoning	256K	Apache 2.0	29	8.1 t/s	on-demand First start 12 s
Qwen 3.8 27B 27B dense (Q4_K_M)	-	ToolsBildReasoning	256K	Apache 2.0	35	19.8 t/s	on-demand First start 76 s
Gemma 4 26B 26B / 4B MoE (Q4_K_M)	-	ToolsBildReasoning	256K	Apache 2.0	19	36.3 t/s	on-demand First start 63 s
Qwen3 Coder Next 80B dense (Q4_K_M)	coder	Tools	256K	Apache 2.0	14	35.3 t/s	on-demand First start 13 s
Qwen3 Coder 31B dense (Q4_K_M)	-	Tools	256K	Apache 2.0	8	48.4 t/s	on-demand First start 10 s
Qwen 3.5 9B 9B dense (Q4_K_M)	-	ToolsBildReasoning	256K	Apache 2.0	14	24.4 t/s	on-demand First start 50 s

15 models · all on dedicated hardware in Upper Austria ^*Independent comparison: Artificial Analysis → Status and availability →

From € 0.25 per million tokens. That's the pricing page.

Input and output priced separately, tiered by model category. 10M tokens free. No base fee, no subscription, usage only.

Category	Input	Output
Flagship	€ 1 / Mtok	€ 4 / Mtok
Reasoning	€ 0.75 / Mtok	€ 3 / Mtok
Multimodal	€ 0.75 / Mtok	€ 3 / Mtok
Code	€ 0.50 / Mtok	€ 2 / Mtok
Standard	€ 0.50 / Mtok	€ 2 / Mtok
Schnell	€ 0.25 / Mtok	€ 1 / Mtok
Privacy detector	€ 0.10 / Mtok	only when active, on input tokens

Prices apply to ALLSmartLLM-owned models. The privacy detector adds only when you use „mask", „pseudonymize" or „strict". External providers (OpenRouter, OpenAI, Anthropic, …) run under your own key. You pay the provider directly, we don't bill anything for them.

Get API key Talk to the person who built it

What we store

This list comes from the log schema, not from a marketing page. Nothing you send through the API is stored: neither your prompts nor the responses.

Field	Via the API	What for
Your prompts	not stored	Never travel from the gateway to a database.
The model's responses	not stored	Delivered, then discarded.
Timestamp	stored	Billing and troubleshooting.
Model requested and used	stored	So you can see what actually answered.
Token counts (in and out)	stored	The basis for billing.
Request duration	stored	Shows when a backend slows down.
Status and error message	stored	Support: without it we cannot help you.
Endpoint called	stored	Billing and abuse detection.
IP address and user agent	stored	Abuse detection and rate limits. 30 days.
Privacy mode selected	stored	Evidence of which level applied to the request.

This applies to every endpoint under /v1/. Storage is on our own hardware in Upper Austria.