Questions, answered.

What is AISynvex?

AISynvex is a multi-model AI platform where models debate, critique, and collaborate rather than just answer. Claude, GPT, and Gemini in one workspace, plus a few flows you don’t get on any single-vendor product: Multi AI Debate (models critique each other across rounds before synthesising), Cowork (specialist Reasoning / Coding / Design roles routed sequentially), and Advanced Thinking (a deliberation layer that adds an arbiter pass on top of any mode — available on every plan).

How does Multi AI work?

The same prompt runs through three models in parallel. On Free, you get all three drafts side-by-side and pick the one you trust — a comparison view, not a fight. On Plus, the models go through debate rounds: each model reads the OTHER two drafts and writes a structured critique, then a synthesizer you pick integrates the strongest insights and the critiques into a single definitive output with attributed influences. The synthesis is informed by what each model thinks the others got wrong, not just by what they independently thought was right.

What is Cowork?

A specialist-role orchestrator. Three fixed roles — Reasoning, Coding, Design — each bound to a model you pick. The Reasoning model gets the first and last say: it runs first to break down the task and decide which downstream specialists are needed, and it runs once more at the end as a Final Review that integrates everything the specialists produced. Coding runs only if the Reasoning step flags requires_code; Design runs after Coding (refining CSS) or directly from the spec when CSS isn’t involved. Skipped roles render as a small pill so the routing decision is transparent. Plus only.

What is Advanced Thinking?

A toggle below the composer that adds a deliberation layer on top of any mode. In Multi AI, an arbiter reviews the synthesis as a final oversight pass (5p per turn). In Single and Cowork, it runs a full four-round flow — three models draft, critique each other, respond to the critiques of their own drafts, then the arbiter reads the entire transcript and writes the definitive verdict (10p per turn). Available on every plan; the arbiter scales with your tier — Free gets Sonnet 4.6, Plus gets Opus 4.6, Pro gets Opus 4.7.

What is AISynvex Forge?

A future autonomous-agent feature, tier-split. Free and Plus get a single-agent Forge: one model takes your task, executes inside a secure sandbox, and escalates blockers back to you for a decision. Pro gets the full three-model debate at every action step — three models propose distinct positions, a synthesizer decides, Cowork-style role routing picks the specialist for each subtask. Both flavours deliver the result (files, output) at the end, not just a transcript. Architecture-pending; no timeline yet.

How does usage work?

Every plan has a rolling 30-day allowance — usage drops out of your window 30 days after each turn rather than all at once at a daily reset. Heavier modes deplete faster: Single is about 1p per turn, Multi AI Comparison about 3p, Multi AI Debate about 7p, Advanced Thinking about 10p. Top-ups are available on every plan when the allowance runs out.

How do top-ups work?

Pick any amount from £1 up to £100 in one click. The four preset tiers: £5 of usage costs £6.49, £10 costs £11.99, £20 costs £22.99, and £50 costs £54.99 — a platform fee is included in each top-up, and higher amounts offer better value. The exact fee for the amount you pick is shown on every purchase before you commit. Top-ups stay valid for 30 days FIFO, kick in automatically once your monthly allowance runs out, and survive plan changes (Plus → Free or vice versa).

Can I attach files?

Yes — every plan, every mode. Text files (txt, md, ts, tsx, js, jsx, py, json, csv, html, css) inject into the model’s system prompt; images (jpeg, png, gif, webp) pass natively as vision content to every model in the call. Limits: 64KB per file, 256KB per conversation total. Works in Single, Multi AI, Cowork, and Advanced Thinking — all read the same attached files.

Is my data used for training?

No. None of the AI providers we route to use API data for training (per their published API terms — Anthropic, OpenAI, and Google all carve their consumer-product training out from API traffic). On our side, your conversations and attachments live in your account so you can return to them; we don’t train any models on your data and we don’t share it with third parties beyond the model providers needed to fulfil your requests. Each provider’s data policy applies to the request payload we forward them; we don’t add our own training pipeline on top.

How does AISynvex compare to ChatGPT, Claude.ai, Poe, or Perplexity?

ChatGPT Plus and Claude.ai Pro are excellent at one vendor’s best models in a polished UI. Poe gives multi-model access but in an aggregator format (one message → one model at a time). Perplexity Max ships a strong autonomous agent (computer use) for hand-off research workflows. AISynvex is multi-model AND multi-mode: the same prompt can run through three models in parallel, the models can debate each other across rounds, or specialist roles can collaborate via Cowork — and Advanced Thinking is on every plan including Free. The side-by-side comparison table below has the detail; we’re honest where the others win.

GDPR + privacy controls?

From the settings page you can: export everything (Article 20 — JSON with conversations, messages, profile, billing history, attached files); delete your account (Article 17 — full cascade, no soft delete). The export is on-demand, generated server-side at request time. Hosting is on Vercel + Supabase (Frankfurt region for EU users). No cookies beyond the auth session and Cloudflare Turnstile bot-prevention.

How does billing work?

Plus is £19.99/month, charged via Stripe. Cancel anytime from the billing page; you keep Plus until the end of the period, then drop to Free. Extra usage top-ups are one-off variable-amount purchases (£1-£100), also via Stripe. No commitments, no exit fees. Pro (£149.99/month) is locked while we finish the team workspace + API features.