R5.112 — initial public release

execai R5.112 — initial public release

execai is a terminal-first AI coding agent that works with your favorite LLM providers (Claude, GPT-5, Kimi K3, GLM-5.2, and more) — bring your own subscription or use ours.

Highlights

9 pluggable LLM sources. ExecAI backend (default) · Z.ai Coding Plan · Kimi Code (kimi.com/code) · Moonshot Platform · Anthropic API · OpenAI API · Claude Code CLI · OpenAI Codex CLI · Ollama (cloud & local). Switch between them on the fly with a shared conversation history.

Ink-style rendering. Message history is committed to the terminal scrollback, so native text selection works and doesn't get reset on stream updates. Live streaming and input live in a dynamic area at the bottom.

Paste collapse. Multi-line pastes become [Pasted #N — L lines, C chars] markers in the input. Full content is expanded when sent to the agent; scrollback stays clean.

Per-source /usage. Real quota display for Kimi Code (weekly + 5h rolling windows), local counters + provider dashboard links elsewhere.

Auto-detect subscription tier. For sources with a /v1/models endpoint (Kimi Code, Moonshot, OpenAI), the agent probes on /connect and shows the actual accessible models tier in the status bar (e.g. kimi (K3 + HighSpeed)).

README in 5 languages — English (primary), Русский, Español, Deutsch, 中文.

Install

Linux / macOS (auto-detects amd64/arm64):

curl -fsSL https://storage.yandexcloud.net/execai-agent-prod/execai/R5/latest/install.sh | bash

Windows 10/11 (amd64/arm64 including Copilot+ PC):

iwr -useb https://storage.yandexcloud.net/execai-agent-prod/execai/R5/latest/install.ps1 | iex

Or download the archive for your platform below, extract, and put on PATH.

Verify

sha256sum -c SHA256SUMS

License

Business Source License 1.1 — free for internal/production use in any company; transitions to Apache 2.0 after 4 years. Commercial hosted/managed service offerings require a commercial license (contact: it@velesbsd.com).

Full documentation

See README.md.