Nyquest Platform Documentation

Nyquest is a multi-model AI workspace. One interface, models from every major provider, transparent live pricing, built-in cost control — plus multi-model consensus, autonomous agents, and media generation in the same workspace.

This page is a tour of the platform at app.nyquest.ai. For step-by-step guides, troubleshooting, an FAQ, a glossary, and the full developer API reference, see the Help Center — it's also built into the app (the ? button, or type /help in chat).

The Workspace & Chassis

The Nyquest workspace is where you interact with AI models — chat, agents, media generation, artifacts, and settings in one place. What makes it different is the chassis system: 11 fully independent workspace designs you can switch between at any time.

A chassis isn't a color theme — each one is a complete design system with its own layout, typography, terminology, and personality. Ridgeline (clinical), Matrix (terminal), Paper (longform writing), Quant (finance, with a live market ticker), Apsis (mission control), Courtroom (legal), Atelier (editorial), Atlas (education, with a built-in Pomodoro timer), Studio (game dev), Garden (wellness), Prism (a look at 2036), and the Nyquest brand chassis. Your conversations, projects, and settings follow you across all of them.

Layout

• Sidebar — conversation history, search, projects, new chat
• Chat area — message input, streaming responses, file upload
• Model selector — pick a model, or leave it on Auto
• Icon rail — Agent Mode, Media Studio, Artifact Vault, Splicer, Settings, Help
• Inspect panel — per-request details (model, tokens, cost, latency)

The workspace is fully responsive — the same features on desktop, tablet, and phone.

Your First Message

Open app.nyquest.ai. Type a message and press Enter. By default the auto-router selects the best model for your prompt, and the response streams back in real time.

Every message creates a conversation with an auto-generated title. Conversations are saved automatically, appear in the sidebar, and survive refreshes and devices.

Try It Without an Account

You don't need to sign up to try Nyquest. Two anonymous paths exist:

• Anonymous free chat — open the app and start typing. Messages are served through Free Mode at $0. No card, no email.
• Anonymous BYOK — paste your own provider API key (OpenAI, Anthropic, Google, OpenRouter, and more), hit Test connection, and chat with that provider's models. The key travels with each request and is never stored, logged, or recoverable — close the tab and it's forgotten.

Signing up adds synced history, memory, projects, artifacts, the full model catalog via wallet funding, and stored (encrypted) BYOK keys.

The Auto-Router

Leave the model picker on Auto and Nyquest routes each message to the model that fits it — by difficulty, by subject matter, and by live price.

Tiers

Every message gets a complexity score that maps to a price tier. Tiers are derived from real, live catalog prices — not model names — so when a model's price changes, its tier updates automatically.

T0 requests stream like any other model and fall through to hosted models automatically if the local tier is unavailable — you never see a degraded reply.

Domain awareness

Beyond difficulty, the router classifies what kind of task you sent — code, creative, reasoning, vision, or general — and picks a model with demonstrated strength in that domain. Routing quality is continuously measured against live traffic, so the ranking reflects how models actually perform, not just their spec sheets.

Provider layer

The same model is often available from more than one upstream provider. Nyquest routes to free capacity first when it exists (the savings show up per-request in your usage data), and providers that misbehave are taken out of rotation automatically. You can bias routing toward a provider you trust with the Preferred Provider setting, informed by live comparisons in Settings → Benchmarks.

Model Pinning

Auto is a default, not a cage. Pick any model from the catalog in the model selector and every message in that conversation goes to it — no routing, no switching. The live per-token price is shown right in the picker before you send.

To go back to automatic routing, select Auto in the model selector.

Supported Models

Nyquest supports models from all major providers, via hosted routing or BYOK:

The catalog is self-maintaining: models that upstream providers list but can't actually serve are detected and hidden from the picker automatically, and restored when they come back. See Models for the live catalog and Pricing for rates.

Free Mode

Anonymous visitors and signed-in accounts with a $0 wallet balance chat for free. Free-mode messages route through a curated chain of genuinely free catalog models, spread across different upstream providers so one outage doesn't take the chain down.

The chain is self-healing — every free model is probed regularly with a real chat-sized request and must return actual content to stay in. Broken models drop out; recovered ones return automatically. Free-mode responses cost you nothing.

Funding your wallet (any amount) or adding a BYOK key moves you off the free chain instantly and unlocks the full catalog.

Multi-Model Splicer

The Splicer sends one prompt to several models in parallel, then judges their answers for consensus. Instead of trusting a single model, you get:

• Agreement % — how much the models' answers align
• Synthesized answer — a single best answer combining the cohort
• Divergences — the specific points where models disagreed, so you know exactly what to double-check

Use it when the answer matters: architecture decisions, fact-heavy research, anything where one model hallucinating would hurt. Your prompt is compressed once before fan-out, so token savings multiply across the cohort — a 4-model splice on efficient models can cost less than a single premium-model reply. Runs are persisted with full per-model detail and costs.

The Splicer is available on funded (Pro) accounts and up — up to 4 models per splice on Pro, 6 on Teams. Read more on the Splicer page.

Agent Mode

Agent Mode takes a goal instead of a message. The agent plans, calls tools, observes results, and iterates — streaming a timeline of step cards so you can watch it work — then lands a final answer in your chat history.

Tools

• Web search — live search for current information
• Web fetch — read a specific page
• Market quotes — live prices for stocks, indices, crypto, commodities
• Project context — your active project's instructions and settings
• Artifact read — pull saved artifacts into the run
• Memory recall — relevant facts from your past conversations

Cost & control

• Billed like chat — per-token, per step. A typical short run costs a few cents to ~$0.50.
• Every run has an automatic cost ceiling based on your recent usage, so a runaway goal can't drain your wallet.
• You can stop a run at any time; completed runs are replayable from history.

Media Studio

Beyond text, Nyquest generates images, voice, music, and video from a prompt — in the same workspace, billed from the same wallet, with no extra provider keys. Every generation shows its exact cost and appears in your billing ledger.

Images

Generate images from a text prompt. Alongside the existing image models, Nyquest offers Google Imagen 4 in three tiers — Fast (~$0.024/image) for quick iteration, Standard (~$0.048), and Ultra (~$0.072) for highest fidelity. Pick a model, choose an aspect ratio, and generate; images land in your gallery.

Voice

Turn text into natural speech. In addition to the standard voices, Nyquest offers 30+ native Google Gemini voices (Kore, Puck, Charon, Aoede and more), with a short style instruction to shape delivery.

Music

Generate original instrumental music with Google Lyria. Describe a mood, instrument, tempo or genre and Lyria composes a ~30-second clip you can play inline and download. Two tiers: Lyria 3 (Clip) and the higher-fidelity Lyria 3 Pro.

Video

Generate short cinematic clips (4–8 seconds, 16:9 or 9:16) with Google Veo 3.1. Video renders as a background job in about 1–2 minutes: the estimated cost is held up front, the panel updates when the clip is ready, and a failed render refunds the hold automatically — you never pay for a clip you didn't get.

Conversations & Memory

Every conversation is stored with full message history and survives across sessions and devices. Auto-generated titles, keyword search, and automatic context management for long threads are built in.

On top of history, Nyquest has a real memory layer: durable facts, rolling summaries, and semantic recall over your past conversations. Relevant memories are surfaced into new conversations automatically, so the platform gets more useful the more you use it — and you can inspect and delete what it remembers at any time in Settings.

Projects

Projects group related conversations under shared context. Each project can carry:

• A system prompt — instructions applied to every conversation in the project
• A preferred model — the default for new conversations in the project
• Its own look — color, icon, and chassis

Drag conversations between projects in the sidebar. Agent Mode picks up the active project's context automatically.

Artifacts

Artifacts are the things worth keeping from your conversations — code, documents, lists, data. Save any response (or fragment) to the Artifact Vault, where it's stored local-first in your browser, searchable by meaning (not just keywords), taggable, and exportable.

Artifacts can be pulled back into any conversation, and Agent Mode can read them during runs.

Web Grounding

Nyquest includes built-in web search. When enabled, live search results are injected into your prompt context before the model sees it — prompts about current events and real-time information get grounded answers. No API keys needed; search runs on Nyquest infrastructure.

Inspect Panel

The inspect panel shows per-request details for every message:

• Model used — which model (and tier) handled the request
• Tokens — input and output counts
• Cost — the exact cost of the request
• Latency — time to first token and total response time
• Routing decision — why this model was selected (when auto-routed)
• Savings — token reduction applied by the optimization engine, and provider savings when served from free capacity

BYOK (Bring Your Own Keys)

BYOK lets you use your own API keys for any supported provider. Your requests go to the provider under your account and billing — Nyquest adds zero markup on BYOK traffic.

Setup

• Go to Settings → Providers, toggle BYOK on for a provider, and paste your key
• Keys are encrypted immediately (AES-256-GCM), decrypted only at dispatch, never logged
• Add or remove keys at any time — OpenAI, Anthropic, Google, xAI, OpenRouter, and more

Prefer not to store the key at all? Use anonymous BYOK — the key is used in-flight per request and never persisted.

Wallet & Billing

Nyquest has no monthly subscription. Pricing is pay-as-you-go from a prepaid wallet:

• Hosted models — requests deduct from your wallet at live per-token catalog pricing with a small, flat platform markup. The picker shows your exact price before you send.
• BYOK models — always free through Nyquest; your provider bills you directly.
• Free Mode and T0 local — $0.
• Fund via Stripe, any amount. Your first funding unlocks Pro: higher rate limits, the Multi-Model Splicer, and live human support.
• When your balance reaches zero, hosted requests pause and BYOK keeps working. No commitments, nothing recurring.

Budget Controls

Optional daily budget controls, enforced automatically:

• Set a daily spending limit at any amount in Settings; usage is tracked in real time per request, per model, per conversation
• When the budget is reached, hosted requests pause until the next day
• Adjust or remove your budget at any time; if no budget is set, usage is limited only by your balance

Agent runs additionally carry their own automatic per-run cost ceiling.

Privacy & Your Data

• Accounts — email + password or Google sign-in; secure token sessions with automatic refresh
• Storage — a usage meter in Settings shows your media storage, with per-file delete
• Memory controls — inspect and delete what the platform remembers about you
• Your data, portable — download everything, delete all conversations, or delete your account entirely from Settings
• Keys — BYOK keys encrypted at rest; anonymous keys never stored; credential headers stripped from logs

Full details in the Privacy Policy and Terms.

Developer API

Everything the web UI does is available over HTTP at api.nyquest.ai. Chat completions are OpenAI-compatible — point any existing OpenAI SDK or tool at Nyquest and it works, with the router and full catalog behind it:

const openai = new OpenAI({
  apiKey: 'nq-pat-...',
  baseURL: 'https://api.nyquest.ai/v1',
});

• Personal API keys — create and revoke in Settings; per-key spend attribution in your usage data
• Streaming — Server-Sent Events on every model
• Beyond chat — endpoints for image, voice, music, and video generation, agent runs, Splicer run detail, live market quotes, billing and usage

The full endpoint reference, auth guide, and rate limits are in the Help Center.