Quick Verdict
OpenAI Sublime Text Plugin is Yaroslav’s community plugin for Sublime Text 4, hosted at yaroslavyaroslav/OpenAI-sublime-text and distributed through Package Control as OpenAI completion. It is not an official OpenAI product and is not affiliated with OpenAI or Sublime HQ. The code is free under the MIT License, but a working setup may still require a Sublime Text license and paid usage from OpenAI, Anthropic, Google, or another remote provider. Ollama and llama.cpp can provide a local backend, with hardware and operational costs replacing the API bill.
It is a sensible choice for committed Sublime Text users who want model chat, selected-code questions, file context, and Build/LSP diagnostics without changing editors. It is not a full agentic IDE. Prompts, selected code, explicitly attached files, and optional diagnostics are sent to the endpoint configured by the user, so retention, training, residency, and contractual protections depend on that provider. Review those terms before using private or regulated source code.
Best For
- Sublime Text 4 users who want AI assistance without moving to another editor.
- Developers who want to switch among OpenAI, Anthropic, Gemini, local models, and compatible endpoints.
- Users comfortable managing API keys, model names, endpoints, budgets, and data policies.
- Developers who want Build or LSP output included in a debugging question.
- Not ideal for teams expecting official OpenAI support, enterprise SLAs, or autonomous repository-wide changes.
Key Features
- Chat using OpenAI Responses, Anthropic, Gemini, and OpenAI-compatible APIs.
- Local model connections through Ollama or a llama.cpp server.
- Project-specific chat histories and assistant settings through
cache_prefix. - Explicit file, tab, selection, Build-output, and LSP-output context controls.
- Chat view and non-destructive inline Phantom output modes.
- SSE streaming plus model, mode, and token information in the status bar.
- Manual copy, append, replace, or open-in-new-tab actions for a response.
Use Cases
Select an unfamiliar function and ask for its contracts and edge cases; attach recent compiler output to investigate a failed build; keep different assistants and histories for separate projects; or point the plugin at an approved local or organizational endpoint. Phantom mode is useful when the user wants to inspect an answer before copying or replacing code. Generated patches still require review, tests, and security checks.
For repository planning and coordinated multi-file edits, compare Cursor and GitHub Copilot. For local inference, Ollama can be used as this plugin’s backend rather than as a direct editor replacement.
Pricing
| Cost item | Position as of July 21, 2026 | Notes |
|---|---|---|
| Plugin | Free, MIT | Community software, no official SLA |
| Sublime Text 4 | Separate license | Not included with the plugin |
| Remote model APIs | Provider pricing | Usage, limits, retention, and availability vary |
| Ollama/llama.cpp | Usually free software | Requires hardware, operations, and model-license review |
The pricing: 免费 field describes the plugin license, not the full workflow. Set provider budgets and rate limits before routine use.
Pros
- Preserves the fast Sublime Text workflow.
- Supports several cloud and local provider types.
- Gives users explicit control over attached context.
- MIT-licensed code can be inspected and modified.
- Phantom output encourages review before integration.
Cons
- Community-maintained and unaffiliated with OpenAI or Sublime HQ.
- More manual configuration than a bundled AI IDE.
- No mature repository-wide planning and agent loop.
- Source-code handling depends on the configured provider.
- Editor licensing, API usage, or local compute may add cost.
Alternatives
| Tool | Best fit | Main difference |
|---|---|---|
| Cursor | AI-first editor users | Deeper repository and multi-file workflow |
| GitHub Copilot | Teams needing broad IDE support | More mature official integrations and governance |
| Ollama | Local-model users | Inference backend that can power this plugin |
| OpenAI Sublime Text Plugin | Existing Sublime Text users | Lightweight, multi-provider, community maintained |
FAQ
Is this an official OpenAI plugin?
No. It is a third-party community project and is not affiliated with OpenAI or Sublime HQ.
Is it completely free?
The plugin is free under MIT. Sublime Text licensing, remote APIs, local hardware, and operations can still cost money.
Where is my code sent?
It goes to the endpoint configured for the selected assistant. That may be OpenAI, Anthropic, Gemini, a third party, or a local service.
Does it support local models?
Yes. It can connect to Ollama, llama.cpp, and compatible APIs. Secure the listener and review each model’s license.
Can it autonomously edit a whole repository?
Its core is chat and controlled context integration. It can append or replace returned text, but it is not equivalent to a full repository agent.
Why was the old slug changed?
sublime-text-ai combined several unrelated extensions and implied a broader official product. The canonical entry now represents this specific plugin; the old slug remains a strict draft and the existing logo is reused.
Bottom Line
OpenAI Sublime Text Plugin adds useful multi-provider chat and diagnostics to Sublime Text 4 without pretending to be an official OpenAI product. Evaluate it with non-sensitive code, verify exactly which endpoint receives context, and budget separately for the editor and model service. Keep it only if that controlled workflow is more valuable than moving to a full AI IDE.