AI privacy

made easy

Discover BlindAI, an open source, accessible and

fast AI inference solution for confidential data.

our mission

Mithril’s mission is to make AI privacy easy. We aim to allow data engineers and developers to train and deploy privacy-friendly AI models with no additional complexity. That is why we are building a ready-to-use Zero-trust middleware for AI inference and training.

our mission

Mithril’s mission is to make AI privacy easy. We aim to allow data engineers and developers to train and deploy privacy-friendly AI models with no additional complexity. That is why we are building a ready-to-use Zero-trust middleware for AI inference and training.

our mission

Mithril’s mission is to make AI privacy easy. We aim to allow data engineers and developers to train and deploy privacy-friendly AI models with no additional complexity. That is why we are building a ready-to-use Zero-trust middleware for AI inference and training.

Protect data at all time

Without BlindAI

With BlindAI

With current solutions, even if data is sent securely,

it is in clear when analyzed by third-party AI solutions.

This means that once you share data externally, you no

longer have control over what is done on it. Malicious

service providers could sell your data, exploit its IP, or

simply get compromised by an attacker & leak your data.

AI user

Exposed Data

AI provider

BLINDAI IN ACTION

BLINDAI IN ACTION

BLINDAI IN ACTION

With BlindAI, leverage state-of-the-art security to benefit from AI predictions without putting users’ data at risk. Even when dealing with sensitive data, AI predictions can be obtained with the guarantee that they have never been disclosed or exposed in clear. Thanks to the hardware protection provided by Intel SGX on top of the software layer we provide, data remains end-to-end protected (at rest, in transit, and use).

We can see what happens before/after BlindAI in the figure above, where a regular AI identification system exposes data, while with our solution, data remains protected.

HOW BlindAI WORKS

import blindai

api_key = "your_api_key" # Enter your API key here

# Upload the ONNX file along with specs and model name
with blindai.Connection(api_key=api_key) as client:
    client.upload_model(”onnx_model_path”, model_id=

”your_model_name”)

Upload your model

Upload your model to a BlindAI secure enclave. By default, you will upload it to BlindAI Cloud, where you need an API key. You can also deploy BlindAI on-premise.

Get predictions with privacy

Send your data to an enclave to have it analyzed, without exposing your data.

import blindai

# Prepare the tensor to be analyzed by the remote AI
input = ...
with blindai.Connection() as client:
    output = client.predict("your_model_name", input)

Step 02
Step 01
AI Teams
Fast & Easy To Deploy
Data Owners

Secure & SEAmless

Compliance Teams

REDUCED LEAKAGE RISK

Join the
community

Join a fast-growing community of developers and

innovators connected all over the world, building

the new era of the internet.

Contribute to our project, and mention open

issues and PRs.

GitHub
Start contributing

Join the community, share your ideas and

talk with Mithril’s team.

Discord
Join the discussion
Open the roadmap

Follow and contribute to our upcoming

projects.

Roadmap
Join the
community

Join a fast-growing community of developers and

innovators connected all over the world, building

the new era of the internet.

Contribute to our project, and mention open

issues and PRs.

GitHub
Start contributing

Join the community, share your ideas and

talk with Mithril’s team.

Discord
Join the discussion
Open the roadmap

Follow and contribute to our upcoming

projects.

Roadmap
Join the
community

Join a fast-growing community of developers and

innovators connected all over the world, building

the new era of the internet.

Contribute to our project, and mention open

issues and PRs.

GitHub
Start contributing

Join the community, share your ideas and

talk with Mithril’s team.

Discord
Join the discussion
Open the roadmap

Follow and contribute to our upcoming

projects.

Roadmap

AI privacy made easy

Discover BlindAI, an open source, accessible and

fast AI inference solution for confidential data.

our mission

Mithril’s mission is to make AI privacy easy. We aim to allow data engineers and developers to train and deploy privacy-friendly AI models with no additional complexity. That is why we are building a ready-to-use Zero-trust middleware for AI inference and training.

Protect data

at all time

With current solutions, even if data is sent securely,

data is clear when analyzed by third-party AI solutions.

This means that once you share data externally, you no

longer have control over what is done on it. Malicious

service providers could sell your data, exploit its IP, or

simply get compromised by an attacker & leak your data.

AI user

Exposed Data

AI provider

Without BlindAI

With BlindAI

BLINDAI IN ACTION

With BlindAI, leverage state-of-the-art security to benefit from AI predictions without putting users’ data at risk. Even when dealing with sensitive data, AI predictions can be obtained with the guarantee that they have never been disclosed or exposed in clear. Thanks to the hardware protection provided by Intel SGX on top of the software layer we provide, data remains end-to-end protected (at rest, in transit, and use). We can see what happens before/after BlindAI in the figure above, where a regular AI identification system exposes data, while with our solution, data remains protected.

HOW BlindAI WORKS

import blindai

api_key = "your_api_key" # Enter your API key here

# Upload the ONNX file along with specs and model name
with blindai.Connection(api_key=api_key) as client:
    client.upload_model("path_to_you_model_in_onnx_format",

model_name="your_model_name")

Upload your model

Upload your model to a BlindAI secure enclave. By default, you will upload it to BlindAI Cloud, where you need an API key. You can also deploy BlindAI on-premise.

Get predictions with privacy

Send your data to an enclave to have it analyzed, without exposing your data.

import blindai

# Prepare the tensor to be analyzed by the remote AI
input = ...
with blindai.Connection() as client:
    output = client.predict("your_model_name", input)

Step 02
Step 01
Compliance Teams

REDUCED LEAKAGE RISK

AI Teams
Fast & Easy To Deploy
Data Owners

Secure & SeAmless

Join a fast-growing community of developers and innovators

connected all over the world, building the new era of the internet.

Join the

community

Contribute to our project, and mention open

issues and PRs.

GitHub

Join the community, share your ideas and

talk with Mithril’s team.

Discord

Follow and contribute to our upcoming

projects.

Roadmap
Open the roadmap

AI privacy made easy

Discover BlindAI, an open source, accessible and

fast AI inference solution for confidential data.

our mission

Mithril’s mission is to make AI privacy easy. We aim to allow data engineers and developers to train and deploy privacy-friendly AI models with no additional complexity. That is why we are building a ready-to-use Zero-trust middleware for AI inference and training.

Protect data at all time

With current solutions, even if data is sent securely,

data is clear when analyzed by third-party AI solutions.

This means that once you share data externally, you no

longer have control over what is done on it. Malicious

service providers could sell your data, exploit its IP, or

simply get compromised by an attacker & leak your data.

AI user

Exposed Data

AI provider

Without BlindAI

With BlindAI

BLINDAI IN ACTION

With BlindAI, leverage state-of-the-art security to benefit from AI predictions without putting users’ data at risk. Even when dealing with sensitive data, AI predictions can be obtained with the guarantee that they have never been disclosed or exposed in clear. Thanks to the hardware protection provided by Intel SGX on top of the software layer we provide, data remains end-to-end protected (at rest, in transit, and use). We can see what happens before/after BlindAI in the figure above, where a regular AI identification system exposes data, while with our solution, data remains protected.

HOW BlindAI WORKS

import blindai

api_key = "your_api_key" # Enter your API key here

# Upload the ONNX file along with specs and model name
with blindai.Connection(api_key=api_key) as client:
    client.upload_model("path_to_you_model_in_onnx_format",

model_name="your_model_name")

Upload your model

Upload your model to a BlindAI secure enclave. By default, you will upload it to BlindAI Cloud, where you need an API key. You can also deploy BlindAI on-premise.

Get predictions with privacy

Send your data to an enclave to have it analyzed, without exposing your data.

import blindai

# Prepare the tensor to be analyzed by the remote AI
input = ...
with blindai.Connection() as client:
    output = client.predict("your_model_name", input)

Step 02
Step 01
Compliance Teams

REDUCED LEAKAGE RISK

AI Teams
Fast & Easy To Deploy
Data Owners

Secure & SeAmless

Join a fast-growing community of developers and innovators connected all over

the world, building the new era of the internet.

Join the community

Contribute to our project, and mention open

issues and PRs.

GitHub

Join the community, share your ideas and

talk with Mithril’s team.

Discord

Follow and contribute to our upcoming

projects.

Roadmap
Open the roadmap

AI privacy made easy

Discover BlindAI, an open source, accessible and

fast AI inference solution for confidential data.

our mission

Mithril’s mission is to make AI privacy easy. We aim to allow data engineers and developers to train and deploy privacy-friendly AI models with no additional complexity. That is why we are building a ready-to-use Zero-trust middleware for AI inference and training.

Protect data at all time

With current solutions, even if data is sent securely,

data is clear when analyzed by third-party AI solutions.

This means that once you share data externally, you no

longer have control over what is done on it. Malicious

service providers could sell your data, exploit its IP, or

simply get compromised by an attacker & leak your data.

AI user

Exposed Data

AI provider

Without BlindAI

With BlindAI

BLINDAI IN ACTION

With BlindAI, leverage state-of-the-art security to benefit from AI predictions without putting users’ data at risk. Even when dealing with sensitive data, AI predictions can be obtained with the guarantee that they have never been disclosed or exposed in clear. Thanks to the hardware protection provided by Intel SGX on top of the software layer we provide, data remains end-to-end protected (at rest, in transit, and use). We can see what happens before/after BlindAI in the figure above, where a regular AI identification system exposes data, while with our solution, data remains protected.

HOW BlindAI WORKS

import blindai

api_key = "your_api_key" # Enter your API key here

# Upload the ONNX file along with specs and model name
with blindai.Connection(api_key=api_key) as client:
    client.upload_model("path_to_you_model_in_onnx_format",

model_name="your_model_name")

Upload your model

Upload your model to a BlindAI secure enclave. By default, you will upload it to BlindAI Cloud, where you need an API key. You can also deploy BlindAI on-premise.

Get predictions with privacy

Send your data to an enclave to have it analyzed, without exposing your data.

import blindai

# Prepare the tensor to be analyzed by the remote AI
input = ...
with blindai.Connection() as client:
    output = client.predict("your_model_name", input)

Step 02
Step 01
Compliance Teams

REDUCED LEAKAGE RISK

AI Teams
Fast & Easy To Deploy
Data Owners

Secure & SeAmless

Join a fast-growing community of developers and innovators

connected all over the world, building the new era of the internet.

Join the community

Contribute to our project, and mention open

issues and PRs.

GitHub

Join the community, share your ideas and

talk with Mithril’s team.

Discord

Follow and contribute to our upcoming

projects.

Roadmap
Open the roadmap

AI privacy made easy

Discover BlindAI, an open source, accessible and

fast AI inference solution for confidential data.

our mission

Mithril’s mission is to make AI privacy easy. We aim to allow data engineers and developers to train and deploy privacy-friendly AI models with no additional complexity. That is why we are building a ready-to-use Zero-trust middleware for AI inference and training.

Protect data at all time

With current solutions, even if data is sent securely,

data is clear when analyzed by third-party AI solutions.

This means that once you share data externally, you no

longer have control over what is done on it. Malicious

service providers could sell your data, exploit its IP, or

simply get compromised by an attacker & leak your data.

AI user

Exposed Data

AI provider

Without BlindAI

With BlindAI

BLINDAI IN ACTION