About Forgecode
ForgeCode is an open source coding harness that runs inside the terminal. It is built on top of Zsh, so typing a colon at the shell prompt opens a conversation with the agent while existing aliases and Oh My Zsh plugins keep working. It connects natively to hundreds of language model providers and models from the shell, and installs through a single setup command published on the project site.
It is aimed at developers who prefer to stay in the terminal rather than move to a separate editor or chat window. A multi agent architecture splits work across specialized sub agents for research, planning and execution, so each step operates on a bounded, relevant slice of context. Models can be mixed inside one session: a reasoning model to plan, a fast model to write code, and a large context model for big files, without restarting. A companion offering called ForgeCode Services adds a context engine for navigating large repositories, rapid tool correction to keep local models on track, and support for large numbers of skills without filling the context window. The vendor states that every change is evaluated across coding tasks and models before it ships, and reports leading results on the TermBench 2.0 benchmark.
No pricing is published. The site has no pricing page, the project is distributed as open source with a public repository and community Discord, and the install path is a shell command rather than a purchase. Running costs in practice come from the model providers you connect, since ForgeCode is the harness rather than the model.
Note that the shell integration assumes Zsh, so developers on other shells should confirm the setup before committing to it.