Talk to your files with AI
Haku brings state-of-art RAG technologies to your premises. Search, summarize, and generate content while keeping your data private. No internet connection needed.
Let information come to you
Haku answers question leveraging all the information in all your file systems.
Haku brings the LLM to you
Haku Outpost
Air gap mode
Haku Compute Outpost
An ecosystem of integrations
Better than all the other options
Build with the cutting edge RAG technology and combined with advanced system design and networking, Haku offers the best user experience in the current market.
Haku V.S. Glean
Haku
Product Offerings
Haku is a software as a server that allows teams to conversationally interact with information in their files that are hosted both in their local file system, storage servers, and cloud storage facilities. Haku focuses on providing the most effertless user experience possible both for accessing information and for operating the software.
IT Overhead
Haku operates completely on-site so there is no need to manage any cloud based resources. Haku also fully automates the setup, operation, and scalling process so only minimum IT knowledge is required to operate Haku.
Scope of Context
Haku focuses on file based knowledge management, so we priortize integration with laptop file systems, file servers, storage servers, and cloud storage facilities.
Glean
Product Offerings
Glean is a cloud native software as a serivce that allows teams to conversationally interact with information in their other cloud native applications. Glean focuses on provideing the widest range of integration into other cloud native applications and services and accessing information on their user's bahalf.
IT Overhead
Glean is a cloud native application that offers fully managed Glean cloud services as well as managed deployment to GCP. However, sophistcated IT knowledge is required to setup and operating Glean on-site.
Scope of Context
Glean focuses on cloud based knowledge management, so we priortize integration with other cloud native applications and services.
FAQs
How does Haku access my file system?
The Haku daemon is responsible for bringing data from your file system into context for Haku. It reads files in the authorized section of your file system and sends the data to the Haku server. The daemon automatically updates the context whenever a file is created, updated, or deleted in the authorized section. All data is encrypted during transit and at rest, ensuring no one, including Haku personnel, can access your data.
What do I need to do to get started with Haku?
Simply sign up for an account (coming soon), download the desktop application, grant Haku permission to access your preferred sections of your file system, and start using Haku. It's that simple.
How can I setup Haku for my team?
After you've signed up for an account, go to the subscription page and select the team plan. Once your subscription has been upgraded, you'll be prompted to create a root identity for your team. Then, create an identity for each team member, install Haku onto their machines, and have them sign in with the credentials you created. That's all there is to it - your team is now set up to use Haku.
What if I cannot let data leave my premise for complience reasons?
For users who can't have their data leave their premises, we offer Haku Outpost, a self-hosted version of Haku available in two options. Users can choose between Haku Outpost on-premise server or Haku Outpost private cloud. Haku Outpost on-premise server involves us sending a physical computing device to your location that can be connected to your network. Haku Outpost private cloud is software that can be installed on your private cloud infrastructure. Both options are designed to keep your data on your premises while still providing the same functionality as the cloud version of Haku.
How does Haku keep all my files in context?
Haku generates an internal knowledge graph based on all the information in your file system. Each file constitutes a section of the graph referencing the path to that file. When Haku answers your question, the Haku AI traverses the graph to find the most relevant information. This allows Haku to answer your questions in context and provide the most relevant information, even from a file you haven't read yet.
How can I make sure the answer Haku generates is accurate?
Since every section of Haku's internal knowledge graph references a file path in your file system, Haku AI uses these file paths to generate a well-formatted answer to your question. This allows you to verify the answer by easily checking the same files Haku used to generate the answer.
