Pricing

Simple plans. Your LLM keys. No token markup.

Try TokenSaver for free, go to production with your own provider keys, and govern AI at the scale of your organisation. You only pay for the platform; your LLM usage stays with your provider, at their prices.

Free

Discover & evaluate

€030-day hosted trial

No credit card

Discover TokenSaver, zero setup.

A hosted trial to test the whole platform on real traffic, with hosted models and no provider key.

  • ✓15 hosted models, no LLM key
  • ✓Cache, RAG, PII and compression included
  • ✓50k requests · 10M tokens / month
  • ✓5k requests / day · 120 per minute
  • ✓3 RAG documents · 1 API key
  • ✓Agent Registry and agentic loops

Pro

Most popular

Agencies, freelancers, SMBs in production

€79excl. VAT / month / user

Monthly · card via Stripe

Go to production with your own keys.

Bring your OpenAI, Anthropic, Mistral… keys, unlock the full catalogue and higher quotas.

  • ✓BYOK: your provider keys, full active catalogue
  • ✓Smart routing
  • ✓100k requests · 25M tokens / month
  • ✓10k requests / day · 200 per minute
  • ✓50 RAG documents · up to 5 GB
  • ✓5 TokenSaver API keys
Get started

Sign up in the console, then upgrade.

Enterprise

IT departments & regulated sectors

Customon contract

Dedicated SaaS or on-premise

Govern security and identity across the organisation.

Enterprise SSO and SCIM, the cyber pack and CCR, on dedicated SaaS or on-premise.

  • ✓Enterprise SSO (OIDC) + SCIM provisioning
  • ✓Cyber pack, CCR and audit
  • ✓500k requests · 100M tokens / month
  • ✓Team memory, agent traces, context graph
  • ✓SLA, dedicated support, on-premise
  • ✓BYOK or dedicated models, full catalogue
Contact sales

Unlimited quotas possible by contract.

01

You pay for the platform

Pro and Enterprise cover cache, RAG, PII, compression and governance.

02

Your LLM, your keys

On Pro and Enterprise, chat runs on your own keys: you pay OpenAI, Anthropic and others directly, at their prices.

03

No token resale, no usage billing

TokenSaver never resells tokens. Pro stops cleanly at its quota, with an alert at 80%.

The Free trial, precisely

30 days of hosted chat. The rest keeps running.

  1. Days 1 to 30

    Hosted TokenSaver chat is included: 15 OpenRouter models, no key needed.

  2. After day 30

    Hosted chat stops. Your Claude egress (your own Anthropic plan), AI flows, policies and loops keep working on Free.

  3. Upgrade to Pro

    Unlocks BYOK chat, more API keys, more RAG and smart routing.

The 15 Free models

  • openai/gpt-5-nano· default
  • openai/gpt-4.1-nano
  • openai/gpt-4o-mini
  • deepseek/deepseek-v4-flash
  • google/gemini-2.5-flash-lite
  • meta-llama/llama-3.1-8b-instruct
  • qwen/qwen-2.5-7b-instruct
  • mistralai/mistral-nemo
  • mistralai/ministral-3b-2512
  • google/gemma-3-4b-it
  • amazon/nova-micro-v1
  • openai/gpt-oss-20b
  • deepseek/deepseek-chat-v3-0324
  • bytedance-seed/seed-2.0-mini
  • z-ai/glm-4.7-flash

Quotas

FreeProEnterprise
Requests / month50,000100,000500,000
Tokens / month10M25M100M
Requests / day5,00010,00020,000
Requests / minute120200300
Responses with tools / month10,00025,000100,000
Tool definitions per request256256512
TokenSaver API keys1550
RAG documents350500
Knowledge storage100 MB5 GB50 GB
Max output tokens (hosted)2,048——

Pro has no unlimited quota: past a limit, calls stop with a 403 and the console warns you at 80%. Enterprise ships with high default ceilings; any quota can be made unlimited by contract.

What’s included

FreeProEnterprise
Semantic cache✓✓✓
Local RAG✓✓✓
Compression✓✓✓
PII detection & anonymisation✓✓✓
Agentic loops✓✓✓
Agent Registry (catalogue, approval)✓✓✓
Your provider keys (BYOK)—✓✓
Smart routing—✓✓
Self-serve checkout (Stripe)—✓—
CCR (recoverable compressed originals)——✓
Cyber pack (policies, posture, alerts, audit)——✓
SSO (OIDC) + SCIM——✓
Team memory, agent traces, context graph——✓
SLA, dedicated support, on-premise——✓
The Enterprise cyber pack coversPrompt injectionDecoratorsTool allowlistOutput handlingContent moderationModel routing

Frequently asked questions

Are LLM costs included?+

No. Pro and Enterprise are platform fees. Chat runs on your own provider keys (BYOK) and you pay OpenAI, Anthropic and others directly, at their list prices. On Free, hosted chat is included for 30 days. RAG and cache embeddings are computed locally: no embedding bill.

What happens after the 30-day trial?+

Hosted chat stops. Egress with your own Claude plan, AI flows, policies and agentic loops keep working on Free. Upgrade to Pro to chat with your own keys.

How do I move to Pro?+

Create your account in the console, then upgrade to Pro: card payment via Stripe, billed monthly.

What happens if I hit a Pro quota?+

Calls stop cleanly with a 403 and the console warns you at 80%. There is no pay-as-you-go overage and no surprise bill.

When does Pro pay for itself?+

As soon as cache, RAG and compression save more than €79 per user per month: typical for support copilots, agents with recurring prompts and document flows. The console shows tokens saved per run.

What does Enterprise add?+

Enterprise SSO (OIDC) and SCIM, the cyber pack, CCR, team memory and agent traces, an SLA and dedicated support, on dedicated SaaS or on-premise, with quotas that can be unlimited by contract.

What is CCR?+

CCR keeps the originals behind compression and lets agents retrieve them on demand, for audit, compliance and regulated sectors. Enterprise only.

Where is my data hosted?+

In Europe. Your LLM keys stay in the TokenSaver vault, never exposed to your apps; your RAG documents stay in your perimeter; traces export to your SIEM for EU AI Act and GDPR audits.

Early Adopter Program

Building something bigger? Let’s talk.

A personalised demo, a governed 30-day POC and Early Adopter terms for teams who want to shape TokenSaver with us.