Private chat across devices
Chat with powerful AI assistants in a private workspace while conversations stay inside secure enclaves. The product is available in the browser and on iOS.
Tinfoil
ClaimTinfoil is a private AI platform for chat, inference, and containerized workloads, keeping data inside secure hardware enclaves across browser, iOS, and API.
Private AI infrastructure and applications
Tinfoil is a private AI platform built around secure hardware enclaves. Its products are aimed at people and teams who want to use AI assistants, inference APIs, and containerized workloads without exposing conversations, requests, or application data outside an attested environment.
The site positions Tinfoil as a way to combine cloud convenience with stronger privacy guarantees. The product line includes Private Chat for private conversations, Private Inference for application builders, and Tinfoil Containers for running Docker-based workloads inside secure enclaves.
Core capabilities
OpenAI-compatible private inference
Build AI applications with an inference API that is OpenAI-compatible, open-source, and fully verifiable. The API is designed to plug into existing OpenAI-based code with minimal changes.
Secure container workloads
Run any Docker image inside a secure enclave with client-side attestation verification. Tinfoil Containers extends the same privacy model to custom workloads and services.
Verified multimodal inference
Get request-level encryption to an attested enclave, plus support for tool calling, vision and image analysis, document processing, and structured outputs.
Private workspace features
Use projects, web search, generative UI, document upload, and encrypted cross-device sync to organize work inside Tinfoil Chat.
Operational controls for containers
Deploy, update, and observe containerized apps with zero-downtime updates, debug enclaves, health checks, and CPU, memory, and GPU monitoring.
Practical ways to use Tinfoil
Private personal or team chat
Use the chat product for sensitive conversations, private brainstorming, or work that benefits from a persistent workspace with projects, document upload, and encrypted sync across devices.
Private AI application development
Build apps that need AI responses without changing existing OpenAI-oriented client code. The inference API supports tool calling, vision, document processing, and structured JSON outputs.
Containerized private workloads
Deploy a Dockerized service, model workload, or internal tool inside a secure enclave when you need client-side attestation and controlled updates.
Privacy-sensitive infrastructure and experiments
Use containers for research or production workflows that need debug environments, private registries, custom domains, and optional GPU support on Hopper and Blackwell hardware.
Pros and Cons
Pros
- Runs AI workloads inside secure hardware enclaves for stronger privacy controls.
- Offers an OpenAI-compatible inference API, which can reduce migration work for existing applications.
- Supports custom Docker images and containerized workloads, not just hosted chat experiences.
- Includes practical operational features such as attestation verification, zero-downtime updates, health checks, and system metrics.
- Provides browser and iOS access for Chat, plus encrypted sync across devices.
Cons
- The public site does not list a full model catalog or every supported integration on the pages provided.
- Some enterprise and researcher options are mentioned for Containers, but pricing and terms are not fully detailed on the public pages.
FAQ
What is the difference between Tinfoil Chat and Tinfoil Private Inference?
Tinfoil Chat runs conversations inside secure enclaves and adds encrypted sync, while the inference API is for building applications that send requests to attested enclaves through an OpenAI-compatible interface.
Is Tinfoil Private Inference compatible with existing OpenAI clients?
The source says the inference API is OpenAI-compatible and works as a drop-in replacement for most existing deployments and workflows.
Can I deploy my own containerized workload on Tinfoil?
Yes. Tinfoil Containers is designed to run any Docker image in a secure enclave, and the site also describes support for private images, debug environments, zero-downtime updates, and health monitoring.
Where can I use Tinfoil Chat?
The site states that Tinfoil Chat is available in the browser and on iOS, and that Secure Sync can keep encrypted backups across devices.
What details are not fully specified on the public site?
The source does not provide a full public list of model providers or pricing tiers for every product, so specific model availability and exact plan details should be confirmed on the relevant product or pricing page.
Quick Facts
- Category
- Private AI platform
- Products
- Private Chat, Private Inference, Tinfoil Containers
- Access
- Browser and iOS for Chat; API and container deployment options for builders
- Core model
- Secure hardware enclaves with client-side attestation verification
- Pricing
- Chat is listed at $20/month; Inference uses usage-based pricing; Containers are priced per hour
- Source domain
- tinfoil.sh