What is Anthara
Anthara runs AI coding agents inside your own network and holds them to your team’s standards. It is built for engineering teams at regulated companies, such as healthtech and fintech, where a change has to meet rules like HIPAA or PCI DSS before it can merge.
Anthara does two jobs:
- It governs the AI coding your team already does. Developers install the
antharaplugin for Claude Code. The plugin loads your organization’s coding standards and shared memory into every session. - It runs agents for you. An agentic workflow is a YAML file that tells an agent what to do and when. A workflow can read tickets, change code in a repository, and open a draft pull request for your team to review.
The parts of Anthara
Section titled “The parts of Anthara”| Part | What it is |
|---|---|
| Web app | The dashboard your team signs in to. Its sidebar is grouped into Assess, Prepare, Connect, Govern, and Operate. |
anthara plugin |
A Claude Code plugin with commands such as /anthara:start and /anthara:plan-implementer. It also connects Claude Code to Fabric. |
| Fabric | The service that stores organization memory and serves your coding standards to AI sessions over MCP. |
| LLM Proxy | A proxy in front of the model provider. Admins use it to register models, set tool rules, protect personal data, and read request logs. |
| Agentic workflows | YAML-defined agents that run on your own cluster. Each step runs in its own short-lived worker. |
How it works explains how these parts fit together.
What you do with it
Section titled “What you do with it”The web app’s sidebar follows the order in which a team usually adopts Anthara.
- Assess. Import a repository and run an assessment. The assessment gives the repository an AI Readiness rating and lists the gaps to fix.
- Prepare. Create AI Coding Standards as packs of rules, and apply the packs to repositories. Review what is stored in Memory.
- Connect. Connect GitHub, Azure DevOps, Jira, and Microsoft Teams so that Anthara can read repositories and tickets and post messages.
- Govern. Register models, decide which tools an agent may call, turn on data protection, and review logs and usage.
- Operate. Create agentic workflows and watch their runs.
Who these docs are for
Section titled “Who these docs are for”- Engineering and platform leaders who decide how AI coding is adopted. Start with How it works.
- Security and compliance reviewers who need to know what is deployed and where data goes. Start with How it works.
- Developers who install and use Anthara. Start with the Quickstart.