Wissly Code launches: an AI coding agent for your internal network

On October 7, 2026, Platum covered Stephow's launch of the AI coding agent Wissly Code. Powered by a local LLM on an internal server, Wissly Code writes and edits code and presents the results for review. It helps organizations work with AI inside their own network when source code and data cannot be sent to external AI services. The original report is in Korean.

Wissly connects company information for search and analysis. Wissly Code focuses on project code, new features, and tools for everyday work.

Describe a task, then review the code and the result

Explain the feature you need, the problem to solve, and the requirements to follow. The agent presents what it will build and check as completion criteria for you to review before work begins.

Wissly Code creates or edits files in the selected project and runs builds and tests in an isolated environment on your internal server. You can inspect file changes line by line and check the running app's appearance and behavior in a preview. Request further changes in the same conversation.

Support development work and everyday tools

Developers can use existing code and documentation to implement features, examine code structure and the effects of a change, or improve comments and technical documentation. The agent can also draft SQL queries and aggregations, CI/CD configurations, Dockerfiles, and deployment scripts.

Business teams can describe inputs and processing rules to create lookup pages, calculators, or scripts for repetitive tasks. A first tool might summarize a CSV by task owner and list overdue items separately. The internal tools AI coding workflow explains how to turn a work process into a tool.

Review queries before applying them to a real database. Production deployment takes place through your internal CI/CD system after reviewing the code and execution results.

Prepare the model and runtime for an isolated network

An environment without internet access needs its components in place before work starts. Review these items for your first deployment:

• Local model: prepare the model files and the environment that runs them on your server.
• Runtime: provide the languages and tools your project needs. The default environment supports Node.js and Python; add other tools as required.
• Packages: provide an internal package source or install the necessary packages in the runtime beforehand.
• First task and review criteria: choose the project, scope, tests, and person responsible for reviewing the result.

See the air-gapped AI coding agent setup guide for deployment preparation. For an existing project, the AI coding agent pilot guide helps define an initial task.

Explore the product screens and workflow on the Wissly Code page.