HomeVortixy APIAPI v1

Vortixy API

Integrate Vortixy Chat into your projects with one Bearer key — chat, images and usage endpoints with the same quotas as your plan. Docs and examples included. Create key in Dashboard → Your API

Same quotas as your plan 00:00 UTC daily reset

Overview

Vortixy is a calm AI workspace for study, writing, research, planning and everyday work. The public API exposes the same Chat you use at /chat so you can call it from your own app, script or workspace. Every key inherits the limits of your current plan — Free, Go $2, Pro $5 or Max $9 — and all quotas are shared between web and API.

Base URL: https://www.vortixy.net/api/v1
Auth: Authorization: Bearer vtx_live_...
Content type: application/json
Quotas: tokens, requests, images, web searches — same as your plan (see Dashboard)

Authentication

Send your key in the Authorization header. Never expose it in client-side code or query params.

curl https://www.vortixy.net/api/v1/chat \
  -H "Authorization: Bearer vtx_live_abc123..." \
  -H "Content-Type: application/json" \
  -d '{"message":"Hello Vortixy"}'

Scopes: chat, images, search, tasks, usage. Keys share your plan quotas.

POST /api/v1/chat

Send a message, optionally include conversation history, enable web search, or adjust the thinking depth. Counts against your plan's daily message and token limits.

Chatbot-ready: add system for your bot persona (8k chars), stream:true for NDJSON deltas, webSearch for live sources. Every call deducts tokens, requests and web searches — shared with web (00:00 UTC reset).

Request fields — all variants

All JSON fields for POST /api/v1/chat. Only message is required — everything else is optional.

Chat request fields
Field	Type	Required	Description
message	string	yes	User message (1–32,000 chars).
history	array	no	Prior turns: [{ role: 'user' \| 'assistant', content: string }] (max 20).
system	string	no	Bot persona / instructions (max 8k chars). Also accepts instructions alias.
provider	string	no	gemini (managed, default). For BYOK, save your key in Dashboard → Settings → AI keys, then use provider openai / anthropic / deepseek / gemini with a BYOK model id.
model	string	no	Model ID — see table below. Must be allowed for provider + your plan.
thinkingLevel	enum	no	none \| low \| medium \| high \| xhigh \| max (xhigh/max only for GPT-5.6).
webSearch	boolean	no	true to enable live web search. Counts toward daily search quota.
stream	boolean	no	true for NDJSON streaming (delta → done). Default false (JSON).

Tip: omit provider/model to use the best model for your plan. Set stream:true for chatbots.

Thinking levels — all variants

Reasoning is now automatic per mode — Fast uses low, Smart uses high, Auto chooses intelligently based on your question. You no longer need to pick a level. The API still accepts thinkingLevel but it is optional and overridden by the mode.

Thinking levels
Level	Description
none	No reasoning — fastest, lowest token use.
low	Light reasoning — quick answers with a little thought.
medium	Balanced reasoning — good for most tasks.
high	Deep reasoning — for complex analysis and planning.
xhigh	Very deep — frontier models only, for hard problems.
max	Maximum — uses the full thinking budget, slowest and most expensive.

Fast → low/medium, Smart → medium/high, Auto → chooses low→high per query. No user selector needed — the system is intelligent per moment.

Chat modes — Fast / Smart / Auto

Pick a mode, not a model. Vortixy routes intelligently to make the most of your quota — no need to choose a specific model. All modes share your plan quotas (00:00 UTC).

Chat modes
Mode	When to use	Thinking	Speed
Fast	Quick answers for simple tasks — fastest, most efficient.	low / medium	Fastest
Smart	Thorough reasoning for complex tasks — deeper, step-by-step.	medium / high	Thorough
Auto	Vortixy chooses the best mode automatically per question (heuristic, zero extra cost).	auto (low → high)	Auto-chooses

Smart has an animated glow when selected (shimmer border). Auto is the default — recommended for most tasks.

BYOK models — 4 providers, free-tier (no plan gate)

Save your provider key once in Dashboard → Settings → AI keys (AES-256-GCM encrypted, never exposed) — then call with provider + model. Same 00:00 UTC quotas as managed.

BYOK models
Model	Provider	Access	Thinking	Context
gemini-3.8-flash:byok	gemini	Free — needs key	none, low, medium, high	1049k
gemini-3.7-flash:byok	gemini	Free — needs key	none, low, medium, high	1049k
gemini-3.6-flash:byok	gemini	Free — needs key	none, low, medium, high	256k
gpt-5.6-sol	openai	Free — needs key	none, low, medium, high, xhigh, max	256k
gpt-5.6-terra	openai	Free — needs key	none, low, medium, high, xhigh, max	256k
gpt-5.6-luna	openai	Free — needs key	none, low, medium, high, xhigh, max	256k
claude-fable-5	anthropic	Free — needs key	none, high	256k
claude-opus-5	anthropic	Free — needs key	none, high	256k
claude-sonnet-5	anthropic	Free — needs key	none, high	256k
claude-haiku-4-5	anthropic	Free — needs key	none, high	200k
deepseek-v4-pro	deepseek	Free — needs key	none, high	256k
deepseek-v4-flash	deepseek	Free — needs key	none, high	256k

BYOK catalog: gpt-5.6-sol/terra/luna (openai), claude-opus-5/fable-5/sonnet-5/haiku-4-5 (anthropic), deepseek-v4-pro/flash (deepseek), gemini-3.8-flash:byok / gemini-3.7-flash:byok (gemini). All available on Free when key is present — no upgrade needed.

Context & automatic compaction

Model context windows vary and reach up to 1M (3.8/3.7) / 256k (rest) tokens. Each request is also limited by your plan: Free 10k, Go 20k, Pro 35k, or Max 50k tokens. In Vortixy Chat (/chat), older turns are automatically compacted above 85% of the applicable budget. The public API (/api/v1/chat) is stateless: store history in your app and resend it in `history` on every call.

We estimate tokens with BPE for system + history + new message.
If total ≤ 85% of the plan and model budget, we keep full history.
If over, we keep the newest turns that fit in 80% of budget, and compact older turns into one summary: “Earlier conversation (automatically compacted…)” with 260-char excerpts.
Recent turns + system + new message are always preserved; the summary keeps useful facts from older turns. In /chat you'll see a `context` event (used/max, compacted, compactedMessages). On the API, you decide what to send in `history` — if you send more than fits (e.g. 200k), the server returns 400/429/502 instead of silently accepting it.

The usable context per request is the lower of your plan limit (10k/20k/35k/50k tokens) and the selected model window (up to 1M (3.8/3.7) / 256k (rest)). For integrations, send only history that fits that applicable budget and summarize older turns yourself; oversized requests may be rejected instead of silently accepted.

Examples — all variants

// Basic — only required field
{ "message": "Explain quantum computing simply" }

// With system persona (chatbot) — 8k chars
{
  "message": "What do you sell?",
  "system": "You are a friendly shop assistant for Acme. Answer briefly in Spanish."
}

// With history — multi-turn conversation
{
  "message": "And what about entanglement?",
  "history": [
    { "role": "user", "content": "Hi" },
    { "role": "assistant", "content": "Hello! How can I help?" },
    { "role": "user", "content": "Explain quantum computing" }
  ]
}

// Full request — all variants together
{
  "message": "Latest AI breakthroughs with sources",
  "history": [{ "role": "user", "content": "Hi" }],
  "system": "You are Acme support. Be concise.",
  "provider": "gemini",          // managed (default) — BYOK dashboard-only for other providers
  "model": "gemini-3.5-flash",    // any available managed model from the table above
  "thinkingLevel": "high",        // none | low | medium | high | xhigh | max
  "webSearch": true,              // true = live web grounding + sources[]
  "stream": false                 // false = JSON, true = NDJSON deltas
}

// 200 OK — non-streaming response
{
  "answer": "Quantum computing uses qubits...",
  "model": "gemini-3.5-flash",
  "usage": { "tokensEstimated": 42, "tokensReserved": 42, "webSearch": true },
  "sources": [{ "title": "Quantum computing — Wikipedia", "url": "https://en.wikipedia.org/wiki/Quantum_computing" }],
  "plan": "pro"
}

// Streaming (NDJSON) — ideal for chatbots
const res = await fetch("https://www.vortixy.net/api/v1/chat", {
  method: "POST",
  headers: { Authorization: "Bearer " + process.env.VORTIXY_API_KEY, "Content-Type": "application/json" },
  body: JSON.stringify({ message: "Hello!", system: "Be concise.", stream: true, thinkingLevel: "medium" })
});
for await (const line of res.body.pipeThrough(new TextDecoderStream()).pipeThrough(splitNDJSON())) {
  const evt = JSON.parse(line);
  if (evt.type === "delta") process.stdout.write(evt.delta);
  if (evt.type === "sources") console.log("sources", evt.sources);
  if (evt.type === "done") console.log("\nusage", evt.usage);
}

// curl variants — pick the one you need
// 1) Basic
curl -X POST https://www.vortixy.net/api/v1/chat -H "Authorization: Bearer $VORTIXY_API_KEY" -H "Content-Type: application/json" -d '{"message":"Hello"}'

// 2) With web search + high thinking
curl -X POST https://www.vortixy.net/api/v1/chat -H "Authorization: Bearer $VORTIXY_API_KEY" -H "Content-Type: application/json" -d '{"message":"Latest AI news","webSearch":true,"thinkingLevel":"high"}'

// 3) Chatbot persona + streaming (NDJSON)
curl -N -X POST https://www.vortixy.net/api/v1/chat -H "Authorization: Bearer $VORTIXY_API_KEY" -H "Content-Type: application/json" -d '{"message":"Hi","system":"You are Acme bot.","stream":true}'

Streaming emits delta → done (NDJSON, Content-Type: application/x-ndjson). With webSearch:true you also get grounded sources (counts against your daily web-search quota).

429 quota exceeded → Retry-After401 invalid key → WWW-Authenticate

GET /api/v1/models

List models your current plan can use (plan-gated, no quota cost). Use it to populate a model picker in your chatbot.

curl https://www.vortixy.net/api/v1/models -H "Authorization: Bearer vtx_live_..."

// 200 OK — example for Pro plan (more models on Max, fewer on Free)
{
  "plan": "pro",
  "defaultModel": "gemini-3.5-flash",
  "models": [
    { "id": "gemini-3.5-flash", "label": "Gemini 3.5 Flash", "provider": "gemini", "thinkingLevels": ["none","low","medium","high"] },
    { "id": "gemini-3.5-flash-lite", "label": "Gemini 3.5 Flash Lite", "provider": "gemini", "thinkingLevels": ["none","low","medium","high"] },
    { "id": "gemini-3.1-flash-lite", "label": "Gemini 3.1 Flash Lite", "provider": "gemini", "thinkingLevels": ["none","low","medium","high"] }
  ]
}

Note: gemini-3.8-flash/gemini-3.7-flash/gemini-3.6-flash temporarily paused (effective 3.5+lite) — example shows effective models only.

Free includes Fast mode only; Go, Pro and Max include Fast, Smart and Auto modes — quotas scale with plan (12k/50k/120k/250k tokens per day, 5/20/50/100 requests). BYOK models need your provider key and are not listed here.

BYOK — use your own provider keys via APIBYOK via API

Bring your own key for OpenAI, Anthropic, DeepSeek or Gemini. Configure once in the app, use everywhere — web and API share the same stored keys and the same daily quotas (00:00 UTC reset).

1. In Vortixy, go to Dashboard → Settings → AI keys and save your provider key (OpenAI sk-…, Anthropic sk-ant-…, DeepSeek sk-…, or Gemini AQ./AIza…). It is encrypted (AES-256-GCM) and never exposed.
2. Call POST /api/v1/chat with provider and model from the BYOK catalog (e.g. provider: "openai", model: "gpt-5.6-terra" or provider: "anthropic", model: "claude-opus-5").
3. Keep using the same Bearer Vortixy key — the provider key is resolved server-side. Quotas, streaming, thinkingLevel and webSearch work the same for managed and BYOK models.

BYOK example

# 1) Save your provider key once in Vortixy (Dashboard → Settings → AI keys)
#    OpenAI sk-..., Anthropic sk-ant-..., DeepSeek sk-..., Gemini AQ./AIza...

# 2) Call the API with the same Vortixy Bearer key, but specify the BYOK provider+model
curl https://www.vortixy.net/api/v1/chat \
  -H "Authorization: Bearer vtx_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Explain quantum computing with deep reasoning",
    "provider": "openai",
    "model": "gpt-5.6-terra",
    "thinkingLevel": "high",
    "stream": false
  }'

# Anthropic example
curl https://www.vortixy.net/api/v1/chat \
  -H "Authorization: Bearer vtx_live_..." \
  -H "Content-Type: application/json" \
  -d '{"message":"Write a haiku about the sea","provider":"anthropic","model":"claude-opus-5"}'

# DeepSeek example
curl https://www.vortixy.net/api/v1/chat \
  -H "Authorization: Bearer vtx_live_..." \
  -H "Content-Type: application/json" \
  -d '{"message":"Solve this algorithm","provider":"deepseek","model":"deepseek-v4-pro","thinkingLevel":"high"}'

// JS — BYOK via Vortixy API (same quotas, same Bearer key)
const res = await fetch("https://www.vortixy.net/api/v1/chat", {
  method: "POST",
  headers: {
    Authorization: "Bearer " + process.env.VORTIXY_API_KEY,
    "Content-Type": "application/json"
  },
  body: JSON.stringify({
    message: "Deep research with my own OpenAI key",
    provider: "openai",
    model: "gpt-5.6-sol",
    thinkingLevel: "xhigh", // xhigh/max only on GPT-5.6 — works via BYOK
    webSearch: true,
    stream: false
  })
});
const data = await res.json();
// data.answer, data.sources, data.usage, data.model

If the key is missing you receive 403 byokMissing — add it in Dashboard and retry. BYOK requests still deduct tokens/requests/searches from your plan.

BYOK catalog: gpt-5.6-sol / terra / luna (openai), claude-opus-5 / fable-5 / sonnet-5 / haiku-4-5 (anthropic), deepseek-v4-pro / flash (deepseek), gemini-3.8-flash:byok / gemini-3.7-flash:byok / gemini-3.6-flash:byok (gemini). Managed Gemini models stay available without BYOK.

POST /api/v1/images

Generate images in standard or quality mode. Counts against your plan's daily image limit.

curl https://www.vortixy.net/api/v1/images \
  -H "Authorization: Bearer vtx_live_..." \
  -H "Content-Type: application/json" \
  -d '{"prompt":"a cozy minimalist living room, warm light","quality":"quality"}'

{ "image": { "mimeType": "image/png", "base64": "..." }, "remaining": 49 }

quality: "standard" (fast, low cost) or "quality" (high detail). Counts against your daily image quota (Free 2/day, Go 8, Pro 30, Max 60).

GET /api/v1/usage

Check your current plan and remaining quotas without spending tokens.

curl https://www.vortixy.net/api/v1/usage -H "Authorization: Bearer vtx_live_..."

{ "plan":"pro","limits":{"aiTokensPerDay":120000,"aiRequestsPerDay":50},"resetsInSeconds": 43200 }

Web Search & Research

Add webSearch: true to let Vortixy fetch live information from the web when it helps. Research mode breaks complex questions into parallel sub-queries and returns cited sources.

await fetch("https://www.vortixy.net/api/v1/chat", {
  method: "POST",
  headers: { Authorization: "Bearer " + process.env.VORTIXY_API_KEY, "Content-Type": "application/json" },
  body: JSON.stringify({ message: "Latest AI breakthroughs with sources", webSearch: true })
});

webSearch: false (default, no web) | true (advanced web search, 5 results, cited sources). Each true call counts 1 toward your daily web-search quota (Free 10/day, Go 25/day, Pro 50/day, Max 100/day).

Tasks & Projects

Stay organized with Projects and scheduled tasks. Manage them from Dashboard or Chat — the API respects the same plan limits for spaces, files, and tasks.

Example: say "Remind me every Monday at 9 am to summarize my week" in Chat — Vortixy creates a scheduled task you can view in Dashboard.

Chats — memoria guardada

Guarda conversaciones en Vortixy para no reenviar el historial cada vez. Mismo límite que tu plan (Free 30, Go 100, Pro 300, Max 400).

GET  /api/v1/chats          # lista tus chats
POST /api/v1/chats          # {title} -> crea uno
GET  /api/v1/chats/:id      # detalle
PATCH /api/v1/chats/:id     # {title} renombra
DELETE /api/v1/chats/:id
GET  /api/v1/chats/:id/messages?limit=50
POST /api/v1/chats/:id/messages  # {role, content}

// POST /api/v1/chats
{ "title": "Mi proyecto Tesis" }
// -> { "chat": { "id": "uuid", "title": "Mi proyecto Tesis" } }

// POST /api/v1/chats/uuid/messages
{ "role": "user", "content": "Hola" }

Files — subir documentos

Sube PDFs, Word, texto o imágenes (20 MiB) para usarlos luego en el chat. Cuenta como artefacto diario (Free 3/día, Go 10/día, Pro 25/día, Max 60/día).

curl https://www.vortixy.net/api/v1/files \
  -H "Authorization: Bearer vtx_live_..." -F file=@tarea.pdf

# JSON alternativo
curl -X POST https://www.vortixy.net/api/v1/files \
  -H "Authorization: Bearer vtx_live_..." -H "Content-Type: application/json" \
  -d '{"name":"tarea.pdf","mimeType":"application/pdf","base64":"..."}'

GET /api/v1/files?limit=20 -> { "files": [{ "id": "...", "name": "tarea.pdf" }] }

Projects — carpetas de estudio

Crea espacios con instrucciones fijas. Límites Free 1 / Go 3 / Pro 8 / Max 15.

GET  /api/v1/projects
POST /api/v1/projects  # {title, subject, instructions}
GET  /api/v1/projects/:id
PATCH /api/v1/projects/:id
DELETE /api/v1/projects/:id

Study Generate — Modo Examen

Genera flashcards o quiz desde un texto (Free 2/día, Go 5/día, Pro 12/día, Max 20/día). Luego puedes exportar a Anki.

curl https://www.vortixy.net/api/v1/study/generate \
  -H "Authorization: Bearer vtx_live_..." -H "Content-Type: application/json" \
  -d '{"content":"texto largo de apuntes...","kind":"flashcards","count":10}'

{ "kind": "flashcards", "items": [{ "front": "¿Qué es...?", "back": "Es..." }] }

POST /api/v1/chat/completions — compatible con OpenAI

Mismo chat pero con formato OpenAI: {messages:[{role,content}]}. Ideal para usar cualquier SDK de OpenAI cambiando solo la URL base a https://www.vortixy.net/api/v1.

// OpenAI SDK
import OpenAI from "openai";
const client = new OpenAI({
  baseURL: "https://www.vortixy.net/api/v1",
  apiKey: process.env.VORTIXY_API_KEY
});
const res = await client.chat.completions.create({
  model: "gemini-3.5-flash",
  messages: [{role:"user", content:"Hola"}]
});
console.log(res.choices[0].message.content);

curl https://www.vortixy.net/api/v1/chat/completions \
  -H "Authorization: Bearer vtx_live_..." -H "Content-Type: application/json" \
  -d '{"messages":[{"role":"user","content":"Hola"}],"stream":false}'

API compatible con OpenAI (2026)

Vortixy expone un espejo 100% compatible con el formato de OpenAI Chat Completions (2026). Cambia solo baseURL y tu apiKey — el mismo código funciona con el SDK oficial de OpenAI, Vercel AI SDK, LangChain o cualquier cliente compatible. Streaming usa SSE estándar data: {...} + data: [DONE] (como OpenAI 2026). Las mismas cuotas del plan (tokens/requests/búsquedas) se comparten con /api/v1/chat.

// OpenAI JS SDK — sin cambiar lógica
import OpenAI from "openai";
const vortixy = new OpenAI({
  baseURL: "https://www.vortixy.net/api/v1",
  apiKey: process.env.VORTIXY_API_KEY, // vtx_live_...
});
const res = await vortixy.chat.completions.create({
  model: "gemini-3.5-flash",
  messages: [
    { role: "system", content: "Eres un tutor amable." },
    { role: "user", content: "Explícame fotosíntesis en 3 bullets" }
  ],
  stream: false,
  // web_search / webSearch también soportado (alias):
  // web_search: true
});
console.log(res.choices[0].message.content);
// res.usage => { prompt_tokens, completion_tokens, total_tokens }
// res.sources => [{title, url}] si webSearch:true

// Streaming SSE (igual que OpenAI 2026)
const stream = await vortixy.chat.completions.create({
  model: "gemini-3.5-flash",
  messages: [{ role: "user", content: "Hola" }],
  stream: true,
});
for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content || "");
}

// Vercel AI SDK (2026)
import { createOpenAI } from "@ai-sdk/openai";
const vortixy = createOpenAI({
  baseURL: "https://www.vortixy.net/api/v1",
  apiKey: process.env.VORTIXY_API_KEY,
});
const { text } = await generateText({
  model: vortixy("gemini-3.5-flash"),
  prompt: "Resume este PDF en 5 puntos",
});

curl https://www.vortixy.net/api/v1/chat/completions \
  -H "Authorization: Bearer vtx_live_..." -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3.5-flash",
    "messages": [{"role":"user","content":"Hola"}],
    "stream": false
  }'

# con búsqueda web (alias 2026 tolerante):
# "web_search": true  o  "webSearch": true  -> grounding Tavily + sources[]

# Python
# from openai import OpenAI
# client = OpenAI(base_url="https://www.vortixy.net/api/v1", api_key="vtx_live_...")
# res = client.chat.completions.create(model="gemini-3.5-flash", messages=[{"role":"user","content":"Hola"}])

Nota 2026: formato idéntico a OpenAI — id: chatcmpl-..., object: chat.completion, choices[0].message, usage y streaming data: {choices: [{delta: {content: "..."}}]} + data: [DONE]. Envelope SSE estándar (OpenAI-compatible). 401 sin Bearer, 429 con Retry-After/RateLimit-*.

Webhooks

Recibe avisos en tu servidor cuando algo pasa. Máx 5 por cuenta.

GET  /api/v1/webhooks          # lista
POST /api/v1/webhooks          # {url, events:["chat.completed","task.completed","quota.warning"]}
DELETE /api/v1/webhooks?id=uuid

OpenAPI

Spec completo en GET /api/v1/openapi para importar en Postman, Insomnia o generar SDK.

curl https://www.vortixy.net/api/v1/openapi

Errors

All errors follow RFC 9457 with a localized message. Check the code field for programmatic handling.

{ "error": "You reached the daily quota. Try again after 2h.", "code": "quotaExceeded", "retryAfter": 7200 }

API errors
Status	Code	When
400	validation	Bad JSON or field out of range
401	apiKeyInvalid	Missing or invalid Bearer key
403	planRequired	Model not allowed for your plan
403	byokMissing	BYOK key missing for this provider — save it in Dashboard → Settings → AI keys
429	quotaExceeded	Daily tokens / requests / searches hit
502	aiEngineFailed	Upstream model error

Headers: RateLimit-Limit, RateLimit-Remaining, RateLimit-Reset, Retry-After, WWW-Authenticate on 401.

Quickstart (Node.js)

Copy, paste, and run. Node 18+ has fetch built in — no extra package needed.

// Node 18+ — no install needed
const res = await fetch("https://www.vortixy.net/api/v1/chat", {
  method: "POST",
  headers: {
    Authorization: "Bearer " + process.env.VORTIXY_API_KEY,
    "Content-Type": "application/json"
  },
  body: JSON.stringify({
    message: "Turn these notes into a plan: ...",
    system: "You are a concise assistant.",
    thinkingLevel: "medium",
    webSearch: false,
    stream: false
  })
});
const data = await res.json();
console.log(data.answer);

// Streaming chatbot
const streamRes = await fetch("https://www.vortixy.net/api/v1/chat", {
  method: "POST",
  headers: { Authorization: "Bearer " + process.env.VORTIXY_API_KEY, "Content-Type": "application/json" },
  body: JSON.stringify({ message: "Hello!", stream: true })
});
for await (const chunk of streamRes.body) {
  // parse NDJSON lines: {type:"delta", delta:"..."} -> {type:"done", answer:"..."}
}

// Image
const imgRes = await fetch("https://www.vortixy.net/api/v1/images", {
  method: "POST",
  headers: { Authorization: "Bearer " + process.env.VORTIXY_API_KEY, "Content-Type": "application/json" },
  body: JSON.stringify({ prompt: "minimalist logo for Vortixy", quality: "quality" })
});
const { image } = await imgRes.json();
// image.base64 -> save as PNG

// Python — same API
// import requests
// r = requests.post("https://www.vortixy.net/api/v1/chat",
//   headers={"Authorization": f"Bearer {API_KEY}", "Content-Type": "application/json"},
//   json={"message": "Hello!", "thinkingLevel": "high", "webSearch": True})
// print(r.json()["answer"])

Try in Chat →Get API key