AI Coding Agents · ovh

shai

shai is a terminal-based AI coding agent written in Rust that acts as a pair programming buddy. It supports interactive chat, headless scripting, HTTP server deployment, and shell command assistance, with multi-provider LLM support including OVHCloud and OpenAI.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/ovh/shai

619

GitHub stars

Forks

Rust

Primary language

Apache-2.0

License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	ovh/shai
Owner	ovh
Primary language	Rust
License	Apache-2.0 — OSI-approved
Stars	619
Forks	56
Open issues	38
Latest release	v0.1.10 (2025-11-13)
Last updated	2025-12-18
Source	https://github.com/ovh/shai

What shai is

A Rust-native CLI tool that integrates LLM capabilities via OpenAI-compatible APIs, featuring interactive REPL mode, streaming SSE HTTP endpoints, MCP agent configuration, and shell integration hooks. Provides both stateful and ephemeral agent modes with project context loading via SHAI.md files.

Quickstart

Get the shai source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/ovh/shai.git2cd shai3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Terminal-native pair programming

Developers seeking an always-available coding assistant directly in their shell without context-switching to web interfaces. Interactive mode allows real-time code generation, debugging, and explanation within existing workflows.

CI/CD and headless automation

Integrate code generation or debugging into build pipelines via headless mode (stdin piping, --trace output). Chain multiple shai calls for multi-step code transformations or automated fixes without manual intervention.

Self-hosted LLM microservice

Organizations needing OpenAI-compatible LLM endpoints on-premises or via private cloud. HTTP server mode provides SSE streaming, stateful sessions, and multimodal APIs suitable for internal developer platform integration.

Implementation considerations

LLM provider setup is mandatory: default OVHCloud anonymous mode is rate-limited. Requires valid API key and authenticated session via 'shai auth' before productive use.
Shell integration ('shai on') injects hooks that monitor command output; review implications for secrets in stderr or sensitive command monitoring before deployment.
HTTP server mode supports both ephemeral (stateless) and persistent agent modes; choose based on cost/statefulness tradeoff and expected concurrency.
Project context via SHAI.md files is loaded at runtime; ensure sensitive information (API keys, internal paths) is not committed to version control.
MCP agent configuration requires manual setup in ~/.config/shai/agents/; agent discovery and hot-reloading are not documented.

When to avoid it — and what to weigh

Proprietary LLM model lock-in required — shai is designed for OpenAI-compatible APIs and MCP. If your organization mandates exclusive use of closed proprietary models without compatible endpoints, integration will require custom work outside shai's supported providers.
Production IDE/LSP integration needed — shai is a CLI-first tool without native IDE plugins or LSP server support. Teams requiring tight editor integration (VSCode, JetBrains) with real-time diagnostics should evaluate purpose-built IDE extensions instead.
Complex multi-model orchestration — shai does not natively support request routing across multiple LLMs, fallback strategies, or model-specific optimizations. If workflow requires dynamic model selection or cross-provider failover, shai is not designed for this.
Enterprise audit/compliance controls — Documentation does not detail audit logging, data residency controls, or compliance certifications. Organizations with strict HIPAA, SOC2, or data sovereignty requirements should review security posture before adoption.

License & commercial use

Apache License 2.0 (Apache-2.0). Permissive OSI-approved license permitting commercial use, modification, and distribution with conditions: include license text, state changes, and provide liability/warranty disclaimers. No patent grant restrictions.

Apache-2.0 permits commercial use, deployment, and redistribution. No restrictions on closed-source deployment or proprietary wrapping. Recommend review of terms around linked/bundled LLM services (OVHCloud, OpenAI) which have separate commercial agreements. shai itself imposes no licensing restrictions on commercial applications.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Adequate
License clarity	Clear
Deployment complexity	Low
DEV.co fit	Good
Assessment confidence	High

Security considerations

Security posture not explicitly documented. Considerations: (1) Shell assistant sends last command and terminal output to LLM; audit what is exposed. (2) API key management via 'shai auth'; verify secure storage in ~/.config. (3) HTTP server mode lacks built-in auth; external reverse proxy required for production. (4) MCP OAuth support mentioned but not detailed. (5) No mention of input sanitization, prompt injection mitigations, or LLM output validation. Requires security review before sensitive workloads.

Alternatives to consider

GitHub Copilot CLI

Mature GitHub-native assistant with IDE and shell integration, built-in GitHub auth, and broader adoption. Better for teams already invested in GitHub ecosystem; less flexible for custom LLM providers.

Continue.dev

Open-source IDE plugin supporting multiple LLM providers and local models. Better for IDE-first workflows; lacks standalone CLI or HTTP server mode that shai offers.

Aider

Specialized terminal agent for multi-file code editing and git workflows. More narrowly focused on code modification; shai is broader (chat, debugging, shell assistance).

Software development agency

Build on shai with DEV.co software developers

Review the architecture, security posture, and LLM provider costs with your engineering leadership before production adoption.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

aichat oh-my-pi shadcn-chatbot-kit

Related on DEV.co

Explore the category and the services that help you build with it.

AI Coding Agents AI application development API development Custom software development Talk to DEV.co

shai FAQ

Can shai run with local/open-source LLMs?

Yes, via any OpenAI-compatible endpoint. Ollama or local LLM servers exposing OpenAI API can be configured, but shai does not include bundled models. Configuration requires custom endpoint URL setup.

What data is sent to the LLM provider?

User prompts, conversation history, shell output (if shell assistant enabled), and project context from SHAI.md files. Documentation does not detail data retention policies of configured providers (OVHCloud, OpenAI). Review provider terms for sensitivity.

Is shai suitable for production deployment?

HTTP server mode is available and can be exposed via reverse proxy. Early versioning (0.1.x) and sparse documentation on scalability, monitoring, and reliability suggest caution. Recommend testing in staging before production use of mission-critical workflows.

Does shai store conversation history?

Interactive mode stores history in session. Headless mode with --trace outputs conversation as JSON. Persistent storage mechanism not documented. HTTP server mode has stateful and ephemeral options; data retention depends on agent configuration and provider backend.

Software developers & web developers for hire

Adopting shai is usually one piece of a larger software development effort. As a software development agency, DEV.co provides software development services and web development expertise — pairing senior software developers and web developers with your team to design, build, and operate ai coding agents software in production.

Evaluate shai for your team

Review the architecture, security posture, and LLM provider costs with your engineering leadership before production adoption.

Talk to DEV.co Schedule a Consultation