Home / Applications
Applications

The enterprise AI experience already exists in open source.

Chat, retrieval over your own documents, coding assistants: the applications your employees want are proven open-source products with huge communities behind them. The hard part has always been giving your people governed access to them: identity, budgets, data boundaries, updates, one support line. That is exactly what Freehold delivers. Every project below ships integrated under our app contract, working the day it lands, kept current across your fleet.

Real projects, under their real names.

The catalog is built from the open-source products these communities have already proven, integrated under the Freehold app contract: licenses inventoried and honored, notices preserved, sign-in bridged to your identity, usage metered against budgets, and updates shipped across the fleet by us.

LibreChat

Chat & content

An enterprise chat workspace with the polish of the consumer tools: threads, agents, file uploads, and search, served in your perimeter against the models your company chose.

  • Multi-model conversations, agents, and file analysis in one interface.
  • Runs against hosted models, or frontier providers under your own keys.
  • OIDC sign-in bridges cleanly to the governed gateway.
Why your people will want it
The chat experience your teams keep asking for, with every thread stored on your side of the wall.
MIT license in the catalog

Continue

Coding assistant

An open-source coding agent for VS Code and JetBrains: autocomplete, chat, and edit-in-place, pointed at your governed model endpoint.

  • Lives in the IDEs your developers already use.
  • Autocomplete, chat, and agent modes, each bindable to the model that fits.
  • Code and prompts stay in your perimeter; docs lookups ride a whitelisted, logged egress.
Why your people will want it
Developers get the assistant they have been asking security to unblock, on models your company runs.
Apache-2.0 license in the catalog

AnythingLLM

Knowledge & RAG

A workspace-based retrieval product: load your documents, ask questions, and get grounded answers with citations, permissioned per workspace.

  • Workspaces map naturally onto your groups and their sources.
  • Answers cite the documents they came from.
  • Ingestion and retrieval run entirely inside your perimeter.
Why your people will want it
Ask the policy binder, the contract archive, and the wiki a question, and get an answer that shows its sources.
MIT license in the catalog

RAGFlow

Deep-document retrieval

A retrieval engine built for messy enterprise paperwork: layout-aware parsing of PDFs, tables, and scans, with traceable, grounded citations.

  • Deep document understanding: tables, figures, and scanned layouts survive parsing.
  • Template-based chunking you can inspect and tune per corpus.
  • An API surface other catalog applications can build on.
Why your people will want it
The contracts and reports your teams actually need answers from are exactly the documents it parses best.
Apache-2.0 license in the catalog

Open WebUI

Chat

A widely deployed self-hosted chat interface with one of the largest communities in the space and a deep plugin ecosystem.

  • Rich chat UI with tools, pipelines, and model management.
  • Large ecosystem of community extensions.
  • Its license adds branding conditions to a BSD base; it enters the catalog once our license review clears governed redistribution.
Why your people will want it
A community favorite; we list it under its real terms while the license review runs.
Open WebUI License still in development

LobeChat

Chat & agents

A modern chat workspace with an agent marketplace, plugins, and strong multi-model support.

  • Polished interface with agents and a plugin marketplace.
  • Broad provider support behind one workspace.
  • Its license adds conditions to an Apache 2.0 base; it enters the catalog once our license review clears governed redistribution.
Why your people will want it
The agent-workspace experience, listed under its real terms while the license review runs.
LobeHub Community License still in development

One memory, shared by every app.

What turns a catalog of separate applications into one product: a member can start a thread in chat and pick the same context back up in the coding assistant, because every app in the deployment sits on the same governed, in-perimeter memory store.

Governed once

Curated at the connector layer

Memory is wired once, beneath every app, instead of per-app integrations multiplying with each addition. Adding an app never silently adds a new place your data lives.

Member-owned

Retention, deletion, export

Every connector that touches member data carries the same obligations: retention limits, clean deletion when a member offboards, export on request, and no administrator back-door into member content.

In-perimeter

Even when the model isn't

Member context is held inside your boundary even when the model answering a given request is an external provider reached under your key.

Know exactly what's on your shelf.

Add-on applications are selected at deployment and added as your needs grow. Every app carries a publisher label, so you always know who's behind the code, and the console pairs each app with its observed adoption, so you can see which ones earn their keep.

Publisher labelWhat it meansWhat you should expect
First-party Built by Freehold, for the catalog. We are the vendor and support path, end to end.
Adopted open source A proven open-source application, integrated under the app contract. Licenses inventoried, notices preserved, branding intact where a license requires it. We are your support-of-record and manage the upstream relationship.
Partner-supplied A commercial vendor's application, integrated by us. The vendor stands behind the code from day one; governance, deployment, and support run through us.
Customer-funded An app a customer funded that entered the shared catalog. Built once, maintained for everyone. Funding brings the app into existence, and the shared catalog keeps it maintained.

The catalog grows with you. Need something we don't stock? We build applications that belong in the catalog for everyone, and customer-funded builds are one of the ways it grows. Every app we add is one we'll maintain for every customer after you, and that shared maintenance is what keeps yours current too.

A deliberate choice

Finished beats flexible.

Agent builders, pipeline canvases, and app-authoring surfaces are aimed at a technical persona most of our buyers don't employ. We put that engineering into the apps themselves instead, so what you deploy is complete.

Every app in the Freehold catalog arrives configured, governed, supported, and updated across the fleet without your team doing anything.

For app founders

Have an app buyers want, but deals stall on governance?

If you've built an AI application and your enterprise deals keep stalling on "where does our data go, can this run in our VPC, who audits usage": catalog placement solves that. We integrate your product under the app contract, and it ships with governance, cost control, and every deployment posture already answered.

The partner program →

Pick the apps and we'll do the rest.

Start from the standard library and tell us what your teams need next. We'll come back with a configured deployment, ready to walk through.

Talk to us