No jargon. No assumptions. If something isn't answered here, email us directly.
GALEN is a language compression system for AI. When you speak or type naturally, AI models have to process every word — and you pay for every word. GALEN converts your natural language into a compact structured token that carries the same meaning in far fewer words.
It works for any subject — medical, legal, financial, travel, business, education, or general conversation. Think of it as a universal shorthand that AI understands precisely.
Here are real examples across different domains:
D.med: patient | sex=F age=45 BMI=42 | T2DM | bypass+next C.travel: itinerary Tokyo duration=3d budget=low >list Q.leg: employment.contract.clauses.key >list S.fin: commercial.property.investment.risks >briefSame meaning as the original. 30–70% fewer tokens. Structured so the AI responds with precision rather than guessing from conversational language.
GALEN was created by George Alex MBBS MRCSEd MCh, a surgeon and founder of Innovations and More Ltd. It grew from a simple observation: the way humans naturally speak doesn't map efficiently onto how AI models process language — leading to wasted tokens, imprecise outputs, and unnecessary cost.
The protocol was built to solve that gap universally — starting in medicine and expanding to every domain where humans communicate with AI.
No. GALEN is a universal protocol. It handles any subject or domain — medical, legal, financial, business, education, travel, technology, creative tasks, or general conversation.
Medicine is where GALEN was first developed and tested, because the precision requirements are highest there. But the compression protocol works for everything.
No. You speak or type exactly as you normally would. The compression happens automatically in the background. You never see or interact with the GALEN token directly — unless you want to.
Anything you would normally say or type to an AI model. Examples include:
Questions — "What are the visa requirements for Japan?" → compressed query token sent to AI
Instructions — "Plan a 3-day Tokyo itinerary on a budget" → compressed command token
Data entry — "Female patient, 45, BMI 42, type 2 diabetes, awaiting bypass" → structured data token
Summaries — "Summarise the key risks of a property investment" → compression token
Generation — "Write a professional email to a client about a delay" → generation token
No. Voice dictation software simply converts speech to text. GALEN goes further — it compresses that text into a structured token that an AI model then processes to generate a precise, structured response.
GALEN is the layer between your voice and the AI. It's the compression and structuring step that makes the AI's job faster, cheaper, and more accurate.
GALEN tokens are model-agnostic. They work with any large language model that accepts text input — including OpenAI GPT models, Anthropic Claude, Google Gemini, Meta Llama, and any other model in your pipeline. No model-specific configuration required.
Three steps:
1. Sign up at galenapi.com and receive your API key (format: gk_live_...)
2. Send a POST request to /v1/convert with your text and API key in the header
3. Receive the compressed GALEN token, reduction percentage, and tokens saved
Full quickstart code in JavaScript, Python, and cURL is at galenapi.com. Integration typically takes under an hour.
The /v1/convert endpoint returns a JSON object with three fields:
{ "galen": "C.travel: itinerary Tokyo duration=3d budget=low >list", "reduction": 52, "tokens_saved": 6 }galen — the compressed token to send to your AI model.
reduction — percentage reduction vs original text.
tokens_saved — absolute tokens saved per call.
Under 200 milliseconds. The compression step adds negligible latency — and the savings on your downstream AI model call more than compensate for the added round trip.
Current domain packs: GALEN-MED (medical), GALEN-FIN (finance), GALEN-LEG (legal), GALEN-TECH (technology), GALEN-BUS (business), GALEN-EDU (education).
Custom domain vocabulary packs are available for enterprise customers with specific terminology requirements.
Starter — £99/month · 10,000 conversions included, then £0.01 per extra conversion.
Growth — £499/month · 100,000 conversions included, then £0.005 per extra. All 6 domain packs.
Enterprise — from £2,000/month · Negotiated volume. White-label, custom domain packs, on-premise deployment.
Full pricing at galenapi.com.
For large enterprise customers, GALEN offers pricing based on demonstrated AI cost savings rather than a fixed fee. If GALEN reduces your AI token bill from £20,000/month to £8,000/month — a saving of £12,000 — we charge £1,000–2,000/month. You keep the rest.
The live demo at galenvoice.com is free and requires no login. Speak or type anything and see the GALEN token and full AI output in real time.
For API access with a test key, contact us at admin@innovationsandmore.com.
Yes. Enterprise plans include white-label rights — deploy GALEN under your own brand within your own product. Custom domain vocabulary packs can also be developed for your specific context.
Contact admin@innovationsandmore.com to start the conversation.
GALEN is patent pending. The patent covers the specific token set, syntax, separator system, and application of the protocol as a whole.
The GALEN API processes text to generate structured tokens. Data handling and retention policies are defined at the enterprise level and can be configured for on-premise deployment, private cloud, or zero-retention operation.
For enterprise deployments with specific compliance requirements, contact us directly to discuss your needs.
Email us at admin@innovationsandmore.com. We respond to all genuine enquiries — whether you're an individual user, a developer, or a platform company exploring integration.
Email us directly — or see GALEN working live in two minutes.