Follow X
Try chat

By default, your chat is encrypted on your device and opened only inside a verified enclave. Not even Thinkr can read it.

  • OpenAI
  • DeepSeek
  • Moonshot AI
  • Z.ai
  • Google
  • Meta
  • 6 models

Claude, GPT, Gemini, and more. The provider reads your chat to answer it and keeps nothing.

  • Anthropic
  • OpenAI
  • Google
  • xAI
  • Mistral AI
  • Qwen
  • DeepSeek
  • MiniMax
  • 118 models

Every other model, video included. The provider reads your chat and may keep it for a while.

  • Google
  • OpenAI
  • xAI
  • Black Forest Labs
  • Kuaishou
  • ByteDance Seed
  • Qwen
  • MiniMax
  • 100 models

224 models from 29 makers, in one app.

In Private mode you type , your device verifies the enclave and encrypts your words for it alone , and the answer comes back the same way .

Private · Verified
4 / 4

  1. Your message

  2. Verification

  3. Encryption

  4. Answer

Every model, one chat.

See more

Chat

  • Pick from 224 models by 29 makers, or let Auto pick for you.

    The model picker of the app: Auto with its intelligence slider, then GPT-OSS 120B, DeepSeek V4.1 Flash, Kimi K3, GLM-5.3, Gemma 4 31B, and Llama 3.3 70B.

  • Choose how private each chat is: Private, ZDR, or Open.

    The lane menu of the app: Private, ZDR, and Open, each with who can read the chat.

  • Talk instead of typing. Your voice turns into text inside the enclave.

    The microphone of the message box, with the tooltip Voice input · Voxtral Small 24B.

  • Drop in a PDF or a photo, or search the web.

    The Actions menu of the message box: Add files or photos, Choose a thinker, Web search, Voice replies, and Privacy filter.

Make images, video, and sound.

Create

  • Make an image in any shape with Nano Banana Pro, FLUX 3, or Seedream 5.0 Pro.

    The settings of the Image page: Create or Edit, the model Nano Banana Pro, the shape, and the number of images.

  • Turn a photo or a few words into video with Veo 3.1 or Kling 3.0 Pro.

    The settings of the Video page: Create, Edit, Upscale, or Talking photo, the model Veo 3.1, the shape, the resolution, and the length.

  • Write a song, give a script a voice, or make a sound effect.

    The settings of the Audio page: Music, Voice, or Sound, the model Eleven v4, and the voice Jessica.

Private by default, verified on your device.

See more

Your privacy

  • There is no account to make. Open the app and start.

    The start screen of the app: the title Think in private, the badge Private · Verified, and the message box, with nothing to sign into.

  • Your browser verifies the enclave before your first word leaves the device.

    The Verification Center of the app: the hardware attested by AMD and NVIDIA, and Runtime is isolated, Passed.

  • Pay with blind notes, so Thinkr cannot link a payment to your chats.

    The card Private billing in the sidebar of the app, on the free trial, with the button Open private billing.

  • Back up your chats encrypted, and open them with your passkey.

    The Backup tab of the settings: Back up chats and credit, and Restore a backup with a passkey or a recovery code.

224 models from 29 makers, in one private app. Zero trackers, zero ads, and zero training on your chats.

Pay less for every model.

Input and output cost 25% less than the priciest other private AI provider charges.

See all prices
  • GPT-OSS 120B−25%

    $0.09 / $0.48

    Other providers$0.13 / $0.65

  • DeepSeek V4.1 Flash−25%

    $0.48 / $1.08

    Other providers$0.65 / $1.45

  • Kimi K3−25%

    $3.00 / $15.00

    Other providers$4.00 / $20.00

Input / output prices per million tokens, on 26 Oct 2026.

Hold THKR, earn credit.

Every 1M THKR in your wallet earns a holder voucher of credit each day, paid in blind notes that never link to your wallet.

See more

  • 1 a day

    holder voucher for every 1M THKR

  • $0.40

    of credit in each voucher, able to rise with the market

  • 100K THKR

    is all it takes to start earning

  • 25% more

    credit when you pay with THKR

The same enclaves, for your code.

Inference and Containers bring the private lane to your own apps and agents.

  • Inference

    An OpenAI-compatible API. The SDK verifies the enclave before each private call, and the rest of your code stays the same.

    See Inference

  • Containers

    Run any Docker image inside an enclave, measured for anyone to verify.

    See Containers

Get started. Think in private.

Start a private chat