AI assistant for safe, reliable productivity with code, collaboration, and integrations.
Whelk
$19/moStarting price
0Popularity
About Whelk
Whelk is a private AI platform designed to run entirely on user-controlled hardware without sending data to external servers. It consists of native applications for Mac, iPhone, iPad, Android, Linux, and terminal environments, alongside a server component that routes models and agents while maintaining a private network. Users can interact with models through chat or activate agents for tasks like coding, with all conversations and files stored locally. The platform supports integration with local model providers such as LM Studio, Ollama, MLX, vLLM, and llama.cpp, allowing flexible deployment across personal or team setups. Whelk emphasizes privacy by ensuring no telemetry, no cloud dependency, and no external servers in the data path unless explicitly added by the user. It is marketed toward individuals and organizations seeking secure, self-hosted AI workflows without operational overhead like Docker, TLS, or backups.
Key features
- Native apps for Mac, iPhone, iPad, Android, Linux, and terminal
- Whelk Server for routing models and managing a private network
- Integration with local model providers (LM Studio, Ollama, MLX, vLLM, llama.cpp)
- Chat and agent modes in the same conversation thread
- File diff approvals and lock-screen permission prompts
- Shared roster, presets, and projects for teams
- Unmetered messages and tokens on all plans
- No account creation required for solo use
Use cases
- Running private AI models on personal or team hardware without cloud dependency
- Secure document analysis and summarization for legal or compliance teams
- Local coding assistance with agent workflows and file approvals
Pros
- Runs entirely on local hardware with no cloud dependency
- Native apps available across multiple platforms including mobile devices
- No telemetry, accounts, or data sent to Whelk servers
- Supports integration with local model providers like LM Studio and Ollama
- Team features include shared roster, presets, and one-scan onboarding
Cons
- Requires installation of a server component for team features
- Windows support is listed as 'in testing' only
- Enterprise features are explicitly marked as 'in testing'
- No free tier for teams; only solo use is free
Frequently asked questions about Whelk
What is Whelk?
Whelk is a private AI platform that runs entirely on user-controlled hardware without sending data to external servers. It provides native applications for Mac, iPhone, iPad, Android, Linux, and terminal environments, along with a server component for routing models and agents while maintaining a private network.
Who is Whelk designed for?
Whelk is designed for individuals and organizations seeking secure, self-hosted AI workflows, including firms in finance, healthcare, legal, defense, manufacturing, pharma, science, accounting, energy, media, and enterprise teams.
How does Whelk handle privacy?
Whelk ensures privacy by design, with no telemetry, no cloud dependency, and no external servers in the data path unless explicitly added by the user. Conversations and files are stored locally, and users control the hardware and endpoints their models interact with.
What pricing model does Whelk use?
Whelk offers a free tier for individuals and a paid team tier based on seats. There are no message caps, token meters, or model tier restrictions, and usage is never metered on any tier.
Can Whelk integrate with local model providers?
Yes, Whelk supports integration with local model providers such as LM Studio, Ollama, MLX, vLLM, and llama.cpp, allowing flexible deployment across personal or team setups.
How do I get started with Whelk?
To get started, users can install Whelk Server on their own hardware and point the native apps to their models. The platform offers multiple entry points, including a team's code, a self-run server, or a provider added by the user.
