CLI

Command-line interface for Tinfoil’s secure AI inference API
GitHub: tinfoil-cli

Overview

The Tinfoil CLI provides a command-line interface for making verified HTTP requests to Tinfoil enclaves and validating attestation documents. It supports all major AI inference operations including chat completions, audio transcription, and text embeddings through a unified inference orchestrator.

Installation

Pre-built Binaries

Download the latest release for your OS from the Releases page.

Install Script

You can install tinfoil CLI using our install script. This script automatically detects your operating system and architecture, downloads the correct binary, and installs it to /usr/local/bin.

curl -fsSL https://github.com/tinfoilsh/tinfoil-cli/raw/main/install.sh | sh

If you receive permission errors (for example, if you’re not running as root), you may need to run the command with sudo.

Build from Source

Ensure you have Go installed.
Clone the repository:

git clone https://github.com/tinfoilsh/tinfoil-cli.git
cd tinfoil-cli

Build the binary:

go build -o tinfoil

Command Reference

Usage:
  tinfoil [command]

Available Commands:
  attestation Attestation commands
  audio       Transcribe audio files using Whisper
  certificate Audit TLS certificate against enclave attestation
  chat        Chat with a language model
  completion  Generate the autocompletion script for the specified shell
  embed       Generate embeddings for the provided text input(s)
  help        Help about any command
  http        Make verified HTTP requests
  proxy       Run a local HTTP orchestrator
  tts         Convert text to speech using TTS models

Flags:
  -h, --help          Help for tinfoil
  -e, --host string   Enclave hostname
  -r, --repo string   Source repo
  -t, --trace         Trace output
  -v, --verbose       Verbose output

Use "tinfoil [command] --help" for more information about a command.

Model Examples

Below are specific examples for each supported model. Click on any model to see its configuration and usage example.

Chat Models

DeepSeek V3.1 Terminus

Alias: terminus

tinfoil chat -m terminus -k "YOUR_API_KEY" "Write a function to validate email addresses and test it with examples"

DeepSeek R1

Alias: deepseek

tinfoil chat -m deepseek -k "YOUR_API_KEY" "Why is tinfoil now called aluminum foil?"

Mistral Small 3.1 24B

Alias: mistral

tinfoil chat -m mistral -k "YOUR_API_KEY" "Explain machine learning concepts"

Llama 3.3 70B

Alias: llama

tinfoil chat -m llama -k "YOUR_API_KEY" "What are the benefits of renewable energy?"

GPT-OSS 120B

Alias: gpt-oss

tinfoil chat -m gpt-oss -k "YOUR_API_KEY" "Analyze database architecture trade-offs for e-commerce"

Qwen3 Coder 480B

Alias: qwen3-coder

tinfoil chat -m qwen3-coder -k "YOUR_API_KEY" "Review this codebase and suggest improvements"

Qwen 2.5 72B

Alias: qwen

tinfoil chat -m qwen -k "YOUR_API_KEY" "Help me debug this Python code"

Audio Models

Whisper Large V3 Turbo

Alias: whisper

tinfoil audio -m whisper -k "YOUR_API_KEY" -f path/to/audio/file.mp3

Kokoro

Alias: tts

tinfoil tts -k "YOUR_API_KEY" "Hello, this is a test of text-to-speech synthesis"

Embedding Models

Nomic Embed Text

Alias: embed

tinfoil embed -k "YOUR_API_KEY" "This is a text I want to get embeddings for"

Chat

The chat command lets you interact with a model by simply specifying a model name and your prompt. You need to specify the model with the -m flag.

Using the Chat Command

Basic Usage (running DeepSeek R1)

tinfoil chat -m deepseek -k "YOUR_API_KEY" "Why is tinfoil now called aluminum foil?"

You can use either the friendly name (deepseek) or the full name (deepseek-r1-0528).

Response Modes

Non-streaming (default): The complete response is returned all at once after generation is finished
Streaming (-s flag): Tokens are displayed in real-time as they’re generated, providing a more interactive experience

# Non-streaming (default)
tinfoil chat -m deepseek -k "YOUR_API_KEY" "Explain quantum computing"

# Streaming response
tinfoil chat -m deepseek -k "YOUR_API_KEY" -s "Explain quantum computing"

Specifying a Custom Model

You can use any model name directly. For models requiring custom enclave settings, supply the -e and -r overrides:

tinfoil chat -m custom-model -k "YOUR_API_KEY" "Explain string theory" \
  -e custom.enclave.example.com \
  -r cool-user/custom-model-repo

If you omit -e or -r for a model that isn’t in the configuration, a warning will be displayed prompting you to specify these flags.

Command Options

-m, --model: The model name to use for chat. Must be specified.
-k, --api-key: The API key for authentication.
-s, --stream: Stream response output (real-time token generation). Optional, defaults to false.
-l, --list: List available chat models.
-e, --host: The hostname of the enclave. Optional if defined in the config file.
-r, --repo: The GitHub repository containing code measurements. Optional if defined in the config file.

Audio

The audio command allows you to transcribe audio files using Whisper models.

Basic Usage

# Using default Whisper model
tinfoil audio -k "YOUR_API_KEY" -f path/to/audio/file.mp3

# Using friendly alias
tinfoil audio -m whisper -k "YOUR_API_KEY" -f path/to/audio/file.mp3

Specifying a Custom Model

tinfoil audio -m custom-whisper-model -k "YOUR_API_KEY" -f path/to/audio/file.mp3 \
  -e custom.enclave.example.com \
  -r cool-user/custom-model-repo

Command Options

-m, --model: The model name to use for transcription. Defaults to whisper-large-v3-turbo.
-k, --api-key: The API key for authentication.
-f, --file: The audio file to transcribe.
-e, --host: The hostname of the enclave. Optional if defined in the config file.
-r, --repo: The GitHub repository containing code measurements. Optional if defined in the config file.

TTS (Text-to-Speech)

The tts command allows you to convert text to speech using TTS models. By default, it uses the kokoro model.

Using the TTS Command

Basic Usage

tinfoil tts -k "YOUR_API_KEY" "Hello, this is a test of text-to-speech synthesis"

This command uses the default model kokoro and saves the generated audio to output.mp3. You can also use the friendly name tts:

tinfoil tts -m tts -k "YOUR_API_KEY" "Hello world"

Specifying Voice and Output File

tinfoil tts -m kokoro -k "YOUR_API_KEY" --voice "af_sky+af_bella" -o "my_speech.mp3" "Custom text to speak"

Command Options

-m, --model: The model name to use for TTS. Defaults to kokoro.
-k, --api-key: The API key for authentication.
--voice: Voice to use for synthesis. Defaults to af_sky+af_bella.
-o, --output: Output file path. Defaults to output.mp3.
-e, --host: The hostname of the enclave. Optional if defined in the config file.
-r, --repo: The GitHub repository containing code measurements. Optional if defined in the config file.

Embed

The embed command allows you to generate embeddings for text inputs.

Basic Usage

# Using default embedding model
tinfoil embed -k "YOUR_API_KEY" "This is a text I want to get embeddings for."

# Using friendly alias
tinfoil embed -m embed -k "YOUR_API_KEY" "Text to embed"

With Multiple Text Inputs

You can provide multiple text inputs to get embeddings for all of them:

tinfoil embed -k "YOUR_API_KEY" "First text" "Second text" "Third text"

Specifying a Custom Model

tinfoil embed -m custom-embed-model -k "YOUR_API_KEY" "Text to embed" \
  -e custom.enclave.example.com \
  -r cool-user/custom-model-repo

Command Options

-m, --model: The model name to use for embeddings. Defaults to nomic-embed-text.
-k, --api-key: The API key for authentication.
-e, --host: The hostname of the enclave. Optional if defined in the config file.
-r, --repo: The GitHub repository containing code measurements. Optional if defined in the config file.

Attestation Verification

Verify Attestation

Use the attestation verify command to manually verify that an enclave is running the expected code. The output will be a series of INFO logs describing each verification step. Sample successful output:

$ tinfoil attestation verify \
  -e inference.tinfoil.sh \
  -r tinfoilsh/confidential-inference-proxy
INFO[0000] Fetching latest release for tinfoilsh/confidential-inference-proxy 
INFO[0000] Fetching sigstore bundle from tinfoilsh/confidential-inference-proxy for digest f2f48557c8b0c1b268f8d8673f380242ad8c4983fe9004c02a8688a89f94f333 
INFO[0001] Fetching trust root                          
INFO[0001] Verifying code measurements                  
INFO[0001] Fetching attestation doc from inference.tinfoil.sh 
INFO[0001] Verifying enclave measurements               
INFO[0001] Public key fingerprint: 5f6c24f54ed862c404a558aa3fa85b686b77263ceeda86131e7acd90e8af5db2 
INFO[0001] Remote public key fingerprint: 5f6c24f54ed862c404a558aa3fa85b686b77263ceeda86131e7acd90e8af5db2 
INFO[0001] Measurements match  

JSON Output

You can also record the verification to a machine-readable audit log:

tinfoil attestation verify \
  -e inference.tinfoil.sh \
  -r tinfoilsh/confidential-inference-proxy \
  -j > verification.json

The audit log record includes the timestamp, enclave host, code and enclave measurement fingerprints, and the verification status.

Certificate Audit

The certificate audit command verifies that a TLS certificate matches the enclave’s attestation document. This ensures the certificate you’re connecting to actually belongs to the attested enclave by comparing the certificate’s public key with the attestation’s TLS public key fingerprint.

Audit from Server

Connect to a server and audit its TLS certificate against the attestation:

tinfoil certificate audit -s inference.tinfoil.sh

You can specify a custom port:

tinfoil certificate audit -s inference.tinfoil.sh:443

Audit from Certificate File

Audit a previously saved PEM-encoded certificate:

tinfoil certificate audit -c /path/to/certificate.pem

Command Options

-s, --server: Server to connect to for retrieving the certificate
-c, --cert: Path to a PEM encoded certificate file
-v, --verbose: Enable verbose output to see detailed verification steps

The command will:

Extract the attestation document from the certificate’s DNS names (encoded in SAN field)
Verify the attestation document’s authenticity
Compare the certificate’s public key fingerprint with the attestation’s TLS key fingerprint
Confirm they match, ensuring the certificate belongs to the attested enclave
Display the enclave measurements for audit purposes

Proxy

Use tinfoil proxy to start a local HTTP orchestrator that verifies connections and forwards them to the specified enclave:

tinfoil proxy \
  -r tinfoilsh/confidential-inference-proxy \
  -e inference.tinfoil.sh \
  -p 8080

Docker

A docker image is available at ghcr.io/tinfoilsh/tinfoil-cli.

Guides

Tool Calling

Learn how to use function calling capabilities with AI models.

Structured Outputs

Use JSON schema validation for reliable data extraction.

Image Processing

Process images with multimodal AI models.

Document Processing

Upload and process documents securely.

Getting Started

Models

SDKs

Guides

Privacy Verification

Tutorials

Admin API

Confidential Computing

Resources

CLI

​Overview

​Installation

​Pre-built Binaries

​Install Script

​Build from Source

​Command Reference

​Model Examples

​Chat Models

​Audio Models

​Embedding Models

​Chat

​Using the Chat Command

​Basic Usage (running DeepSeek R1)

​Response Modes

​Specifying a Custom Model

​Command Options

​Audio

​Basic Usage

​Specifying a Custom Model

​Command Options

​TTS (Text-to-Speech)

​Using the TTS Command

​Basic Usage

​Specifying Voice and Output File

​Command Options

​Embed

​Basic Usage

​With Multiple Text Inputs

​Specifying a Custom Model

​Command Options

​Attestation Verification

​Verify Attestation

​JSON Output

​Certificate Audit

​Audit from Server

​Audit from Certificate File

​Command Options

​Proxy

​Docker

​Guides

Tool Calling

Structured Outputs

Image Processing

Document Processing

Overview

Installation

Pre-built Binaries

Install Script

Build from Source

Command Reference

Model Examples

Chat Models

Audio Models

Embedding Models

Chat

Using the Chat Command

Basic Usage (running DeepSeek R1)

Response Modes

Specifying a Custom Model

Command Options

Audio

Basic Usage

Specifying a Custom Model

Command Options

TTS (Text-to-Speech)

Using the TTS Command

Basic Usage

Specifying Voice and Output File

Command Options

Embed

Basic Usage

With Multiple Text Inputs

Specifying a Custom Model

Command Options

Attestation Verification

Verify Attestation

JSON Output

Certificate Audit

Audit from Server

Audit from Certificate File

Command Options

Proxy

Docker

Guides