// Features
Everything Gate does for you.
Forty-odd concrete reasons — cost, quality, privacy, control, Odoo and developers — behind one API key.
Top up →See pricing →Forty reasons it pays off.
Pick yours.
Cost
No more overpaying.
Simple tasks automatically go to low-cost models — only the hard questions get the expensive ones.
Balance or plan.
Top up and pay per use with no monthly fee, or take a plan with a fixed monthly price. You get exactly what you pay for.
In your own currency.
You pay and see your spending in your billing currency — not in dollars, converted after the fact.
A proper invoice.
A compliant invoice for every top-up — ready for your bookkeeping, not a card receipt from an overseas provider.
One invoice instead of many.
Instead of several AI providers: one supplier, one balance, one line in your books.
Prompt caching.
Recurring instructions and context are reused behind the scenes — repeated calls are cheaper, with no setup.
Half-price background processing.
What can wait (say, a hundred invoices overnight) runs as a batch at half the normal price.
A limit per key.
Give every key its own ceiling. A runaway automation can never spend your entire balance.
Valid for 12 months.
Your balance can be used for a full year — no need to “use it up” every month.
Quality and reliability
Always the best models.
When a better or cheaper model comes out, we connect and test it. You don’t have to change a thing.
Automatic fallback.
If a provider slows down or goes offline, the request automatically moves to the next model. Work doesn’t stop.
Multiple providers, one key.
You are not at the mercy of a single company’s pricing or outages.
Continuous monitoring.
Response times, error rates and provider health watched around the clock, with alerts.
Full answers, even when it thinks.
With reasoning models you still get an answer of the length you asked for — the thinking doesn’t eat your limit.
Images and documents.
Invoice photos, scanned delivery notes, PDFs — the models can see, not just read.
Speech to text.
Transcription with the same key, from the same balance.
Search and RAG.
An embedding model to make your own documents searchable — on European infrastructure.
No hard-coded decisions.
Where the AI decides, the AI really decides — per request, based on the content, not from a pre-written rule list.
Privacy and security
Runs in Europe.
Gate runs in Frankfurt, on our own infrastructure.
Your secrets never leave.
API keys, tokens and passwords are automatically swapped out of every request — the model only sees a placeholder, and the original value is put back in the answer.
Identifier masking when you need it.
With one switch, e-mail addresses, phone numbers, IBANs, card numbers, and tax and personal ID numbers are masked before they leave — if your internal policy requires it.
No training on your data.
We only use provider APIs that contractually commit not to train on your data.
Customers kept apart.
Caches are isolated per customer — two companies’ requests never mix.
Disable a key in one click.
Lost laptop, departing colleague? Disable that key instantly — everything else keeps working.
A log without the text.
Time, model, cost and status of every call can be looked up — without the content of the conversations.
One data processor.
Instead of several provider contracts, one UK-registered provider is responsible — and your data stays in the EU, in Frankfurt.
GDPR and UK GDPR.
A UK-registered provider with infrastructure in the EU: built to meet both the EU GDPR and the UK GDPR.
Control and transparency
Live balance.
Always see how much you have left — not just at the end of the month.
Statistics per task.
What was the AI used for? Invoice processing, e-mail writing, chat — with counts and costs.
Statistics per model.
Which level and which model did the work, and what it cost.
The last 100 calls.
Time, task, model, response time, cost and status — per key.
A key per colleague.
See exactly who uses how much — with a ceiling, a note and an owner.
Warnings in time.
Get an e-mail before your balance runs out — and see roughly how many days it will last.
eYssen and Odoo
Built in out of the box.
In the AI modules of eYssen (Odoo CE fork), eYssen AI Gate is the default provider — enter your key and you are done.
Tuned for business tasks.
Invoice reading, e-mail drafting, field translation, summaries, chat and agents — all through Gate.
Statistics inside your system.
Balance, expiry, usage and the latest calls right on the provider settings.
“Can wait” processing.
AI fields can also be filled in the background, as a half-price batch.
At home in tax and accounting.
19 years of Odoo experience built into the AI features — including tax, accounting and localisation questions.
Not just eYssen.
The key works in any system — eYssen is simply the most convenient place to use it.
For developers and teams
OpenAI-compatible API.
Any tool, library or agent framework that speaks OpenAI works with Gate — just change one URL and one key.
Anthropic format too.
The Messages API is available — Claude-based tools run on the same key.
One key for the team.
Developers don’t each need their own account with several providers.
Long-term stability.
The model name stays the same (eyssen/auto) while the models behind it are kept up to date — your code doesn’t change.
5 minutes.
1 key.
Every model.
Top up your balance, create a key, and today the right AI for each task is working for you — in eYssen or anywhere else.
Top up your balance →Talk to us