April 2026 · 37 models · £0/month

Your work,
amplified.

Memo AI — an intelligent work assistant built internally for Memo Fashion. Write emails, analyse documents, generate images, search the web, and more.

or

Only @memo.co.uk accounts are permitted

Online
Good morning, Memo
How can Memo AI help you today?
✉️ Write an email
🔍 Search the web
📊 Analyse data
💡 Brainstorm ideas
Ask Memo AI anything…
→
Production · April 2026

Built in-house.
37 models. Zero cost.

Memo AI runs on 37 live-verified AI models across 7 providers — Groq, Cerebras, SambaNova, Gemini, OpenRouter, Cloudflare and Tavily. Every model cascades automatically: if one fails, the next fires in under 50ms. Chat history is encrypted and stored in Supabase. File attachments persist in Cloudflare R2. Total monthly cost to Memo Fashion: £0.

37
Live AI models
32
Rotating API keys
7
Free-tier providers
£0/mo
Total running cost

Made for you.

Memo AI is built by the Memo internal team for the people of Memo Fashion. Here's what we made you.

✨
Pick your brain
Six AI models — Smart (DeepSeek V3.2), Reasoner, Live (Gemini + Google Search), Fast (41ms), Coder, Vision. Each backed by 7-12 free-tier engines that cascade automatically.
→
⚙️
Under the hood
Architecture, cascade design, 37-model matrix, latency benchmarks, FLUX.2 klein image editing, and how Memo AI compares to ChatGPT, Claude and Gemini.
→
🛤
The journey
17 months of work compressed into a timeline. Every release, every breakthrough — from rough prototype to the version you're using right now.
→
🛠
Built in-house
By the Memo Internal Software Team. Engineered, refined and shipped by people who actually work alongside you. Not a wrapper. Not outsourced. Yours.
→
Six models · one chat

Pick the brain for the job.

Memo AI isn't one model — it's six, each wired to do something different brilliantly. Switch between them mid-chat. No extra sign-up. No limits. Here's what each one does and what powers it underneath.

✨
Smart
Default · everyday
The all-rounder. Polished emails, summaries, analysis, brainstorming, translation. Picks the fastest free model available in real time — you never wait.
Powered bySambaNova DeepSeek V3.2 · Groq GPT-OSS 120B · SambaNova Llama 4 Maverick · Cerebras Qwen 3 235B · OpenRouter Nemotron 120B · GLM-4.5 Air — cascade of 12 free models.
🧠
Reasoner
Deep think
For hard problems that need step-by-step logic. Complex code, multi-step maths, strategy, legal reasoning. Slower than Smart, but thinks before it speaks.
Powered bySambaNova DeepSeek V3.2 · DeepSeek V3.1 · Cerebras Qwen 3 235B · Groq GPT-OSS 120B · OpenRouter Nemotron 120B. Reasoning traces visible on demand.
🔴
Live
Real-time
Actually knows what happened today. Current news, weather, prices, sports scores. Grounds every answer in live Google search results, with sources cited at the bottom.
Powered byGemini 2.5 Flash with Google Search grounding · Groq GPT-OSS 120B + Tavily fallback. 9 rotating Gemini keys for unlimited day-to-day use.
⚡
Fast
2000 tokens/sec
Quick questions, quick answers. One-liner lookups, “what's the word for…”, quick rewrites. Routed to Cerebras's wafer-scale chips — the fastest inference hardware on earth.
Powered byGroq GPT-OSS 20B (41ms first token) · Groq Llama 3.1 8B Instant · Cerebras Llama 3.1 8B (WSE-3 chip) · OpenRouter Nemotron Nano 9B fallback.
💻
Coder
New · programming
Tuned for code. Generates clean TypeScript / Python / SQL, explains patterns, spots bugs, writes tests. DeepSeek V3.2 is top-tier for refactoring; GLM-4.5 Air is free-tier fallback.
Powered bySambaNova DeepSeek V3.2 (best for code) · Groq GPT-OSS 120B · Cerebras Qwen 3 235B · OpenRouter GLM-4.5 Air :free · Qwen3 Coder :free.
👁
Vision
Auto · attach any image
Reads screenshots, photos, receipts, handwriting, diagrams, charts. Switches on automatically the moment you attach an image — no toggling. Plus: paste an image and say “make this burgundy” to edit it via free Cloudflare img2img.
Powered byGroq Llama-4-Scout 17B · Gemini 2.5 Flash · OpenRouter Gemma 3 27B · Cloudflare FLUX.2 klein 9B/4B for instruction-following edits.