Devgrex
devgrex.com
- Free plan
- Windows
- macOS
- Linux
- Web
- API
- Self-hosted
What Devgrex is
Devgrex is an AI-native development environment combining a code editor, autonomous agents, and model routing. Its Tab model provides context-aware multiline completions with partial acceptance and ghost-text previews; local-model latency is listed as under 100ms. Agents can edit multiple files, run terminal commands in a sandbox, detect and fix errors, and manage Git operations. Cloud Agents run tasks on a cloud VM and can create GitHub or GitLab repositories, push branches, and open pull requests. Developers can use managed models from Anthropic, OpenAI, Google, and Moonshot, bring their own provider keys, or connect local models through Ollama and compatible servers. The environment also offers project-wide semantic code search, MCP connections to external tools, and compatibility with VS Code extensions, themes, keybindings, and settings. Privacy options include local completions and offline operation, with automatic redaction before cloud requests. The free plan includes local completions, agent chat, and offline mode. Devgrex was founded in 2025 and is based in Tallinn, Estonia.
Who it suits
Devgrex suits developers looking for editor-based AI assistance, agent workflows, or a choice of hosted and local models. Its offline mode and enterprise security options may also fit teams with stricter privacy or access requirements.
Good to know, and what to watch
- Tab offers multiline completions and partial acceptance.
- Agents can edit files and manage Git operations.
- Cloud Agents can open GitHub or GitLab pull requests.
- Supports local models and bring-your-own provider keys.
- Free plan includes offline mode.
- Cloud agent tasks run on a cloud VM.
- Some model use involves cloud requests, though privacy modes and redaction are available.
Plans
Every plan and priceFree
Free plan
Getting started
- Open the Devgrex website
- Choose a supported desktop, web, API, Linux, macOS, Windows, or self-hosted route
- Create an account and select the Free plan or another available arrangement
- Connect a managed model provider, bring your own key, or configure a supported local server
- Open a project to use semantic indexing and search
- Install or apply compatible VS Code extensions, themes, keybindings, or settings
Limits to check
The Free plan includes local completions, agent chat, and offline mode. The listed facts do not define additional quotas or seat limits for that plan.
The verdict
Devgrex combines code completion, agent tasks, model selection, and project-wide search in one development environment. Its free plan includes local completions and offline mode, while teams can also use its listed enterprise security features.
Questions people ask
Which platforms and deployment options does Devgrex support?
It supports API, Linux, macOS, self-hosted, web, and Windows use.
Is there a free plan?
Yes. The Free plan costs nothing and includes local completions, agent chat, and offline mode.
Which model providers can I use?
Managed models include Anthropic, OpenAI, Google, and Moonshot. BYOK is supported for every provider, and local connections include Ollama and compatible servers.
Does Devgrex support VS Code extensions?
Yes. Its desktop IDE is compatible with VS Code extensions, themes, keybindings, and settings.
What integrations are available?
MCP can connect tools such as Figma, Slack, databases, APIs, documentation servers, Jira, and Confluence.
What enterprise security features are listed?
Enterprise features include SAML 2.0 and OpenID Connect SSO, role-based access, audit logs, VPC deployment, and air-gapped deployments with local models.
What the maker publishes
- Agent mode
- Agents edit multiple files, execute terminal commands in a sandbox, detect and fix errors, and manage Git operations. source
- API
- The initial release included a REST API v1 covering chat, completions, model catalog, account functions, and billing webhooks. source
- Autocomplete
- Its Tab model provides sub-100ms local-model latency, context-aware multiline completions, partial acceptance, and ghost-text previews. source
- Cloud agents
- Cloud Agents run tasks on a cloud VM, can create GitHub or GitLab repositories, push branches, and open pull requests. source
- Code intelligence
- The product provides semantic codebase indexing and search across an entire project. source
- Enterprise security
- Enterprise features include SAML 2.0 and OpenID Connect SSO, role-based access, audit logs, VPC deployment, and air-gapped deployments with local models. source
- Extensions
- The desktop IDE is compatible with VS Code extensions, themes, keybindings, and settings. source
- Founded
- 2025 source
- Headquarters
- Tallinn, Estonia source
- Integrations
- MCP support connects external tools such as Figma, Slack, databases, APIs, documentation servers, Jira, and Confluence. source
- Local models
- Devgrex supports Ollama and OpenAI-compatible servers including LM Studio, llama.cpp, vLLM, Jan, and LocalAI. source
- Model providers
- Managed models include Anthropic, OpenAI, Google, and Moonshot models, with BYOK support for every provider. source
- Privacy controls
- Privacy modes include Normal, Sensitive with local completions, and Offline air-gapped operation; automatic PII and secret redaction is available before cloud requests. source
- Product
- Devgrex is an AI-native development environment combining a familiar editor, autonomous agents, and model routing. source
- Security
- Devgrex uses TLS 1.2 or higher for API traffic, encrypts account data at rest, and says its cloud API does not store source code or prompt content. source
- Support
- Documentation directs users to a community Discord and GitHub Issues for help. source
Checked against the maker's pages on 1 Oct 2026.