Assistant not working in editor (but works in console and sidebar)

[fdetsch]

Hi all,

Following @juliasilge’s suggestion, I'm turning my comment on a previously closed issue into a discussion.

I'm currently experiencing an issue where the Positron assistant works in the console and sidebar, but not in the editor. The original issue #12573 appears to be closed, so I wanted to check whether this behaviour has been resolved or if this is still a known problem.

I'm attaching the Positron assistant diagnostics for reference.

Happy to provide additional details if needed.

Thanks!

Positron assistant diagnostics:

# Positron Assistant Diagnostics

Generated: 2026-04-15T18:51:14.408Z

**Privacy Notice**: This diagnostic report includes:
- Extension versions and configuration settings
- Model configurations (including base URLs and model IDs)
- System information (OS, architecture)
- Recent log entries
- Chat session metadata

**The report does NOT include API keys or authentication tokens.** However, base URLs may reveal internal endpoints, and configuration settings might expose security policies. Please review carefully before sharing.

## Version Information

- Positron Assistant: Version 0.0.1 (Active)
- GitHub Copilot Chat: Version 0.37.5 (Active)
- Positron: 2026.04.0 (build 269)
- Code OSS: 1.109.2
- Application: Positron
- OS: linux x64
- Remote: wsl

## Configuration Settings

### Extension Settings

Positron Assistant and GitHub Copilot settings:

```json
  "positron.assistant.enable": true,
  "positron.assistant.inlineCompletions.enable": {
    "*": true
  },
  "positron.assistant.aiExcludes": [
    "**/.*"
  ],
  "positron.assistant.inlineCompletionExcludes": [
    "**/.*"
  ],
  "positron.assistant.maxInputTokens": {},
  "positron.assistant.maxOutputTokens": {},
  "positron.assistant.models.include": [],
  "positron.assistant.notebookSuggestions.model": [
    "haiku",
    "mini"
  ],
  "positron.assistant.providerVariables.bedrock": {},
  "positron.assistant.providerVariables.snowflake": {},
  "positron.assistant.models.preference.byProvider": {},
  "positron.assistant.models.overrides.anthropic": [],
  "positron.assistant.models.overrides.amazonBedrock": [],
  "positron.assistant.models.overrides.snowflakeCortex": [],
  "positron.assistant.models.overrides.openAI": [],
  "positron.assistant.models.overrides.customProvider": [],
  "positron.assistant.models.overrides.positAI": [],
  "positron.assistant.models.overrides.msFoundry": [],
  "positron.assistant.models.overrides.google": [],
  "github.copilot.nextEditSuggestions.enabled": true,
  "github.copilot.enable": {
    "*": true,
    "plaintext": false,
    "markdown": false,
    "scminput": false
  },
  "github.copilot.chat.reviewSelection.instructions": [],
  "github.copilot.chat.githubMcpServer.toolsets": [
    "default"
  ],
  "github.copilot.chat.codeGeneration.instructions": [],
  "github.copilot.chat.testGeneration.instructions": [],
  "github.copilot.chat.commitMessageGeneration.instructions": [],
  "github.copilot.chat.pullRequestDescriptionGeneration.instructions": [],
  "github.copilot.chat.anthropic.tools.websearch.allowedDomains": [],
  "github.copilot.chat.anthropic.tools.websearch.blockedDomains": [],
  "github.copilot.chat.planAgent.additionalTools": [],
  "github.copilot.chat.debugTerminalCommandPatterns": []

Language Model Providers

Configured Providers

No models configured

Available Models

copilot (8 models)

• Claude Opus 4.6 (claude-opus-4.6) - Max Input: 167805
• Claude Sonnet 4.6 (claude-sonnet-4.6) - Max Input: 167805
• GPT-5.4 (gpt-5.4) - Max Input: 271805
• GPT-4o mini (gpt-4o-mini) - Max Input: 12093
• GPT-4o mini (copilot-fast) - Max Input: 12093
• GPT-4o (gpt-4o) - Max Input: 63805
• GPT-4.1 (gpt-4.1) - Max Input: 111424
• Auto (auto) - Max Input: 167805

Active Chat Session

No active chat session

</details>

[jonvanausdeln]

@fdetsch , Thank you for reaching out. Can you provide a few more details?

• Which Positron version and OS are you using?
• Can you send a screenshot of what you see when press ctrl + i in the editor to pull up the inline chat? And give a little more context of what you were trying to accomplish?

Inline assistant is much more limited it what it can do, it doesn't have all the same capabilities as the full chat.

[fdetsch]

@jonvanausdeln While creating the short demo attached below, I realized that the behaviour described above only applies to specific models. My company hadn't enabled higher-level Copilot models until this afternoon, so I could only test with GPT-4 and always ran into this issue.

Anyway, as shown in the video, the sidebar and terminal inline chat work with GPT-4, while the model seems to think for a few seconds in Editor mode before going back to “Generate code” without actually creating anything. With higher-level models, the problem disappears.

I added my Positron assistant diagnostics in a collapsible section above. The Positron version and OS (Ubuntu 24.04 via WSL2 running on Windows 11) are included there.

Here's the aforementioned video:

Editor.In-Line.Chat.Problem.with.GPT-4.mp4

[jonvanausdeln]

Thank you @fdetsch for the new detailed information.

I've filed an issue so you can track progress there or add any additional information. #13065

[HarmHartman]

Hi, I was running into the same issue. For some reason this setting was set to false on a new installation. Setting this to true fixed it.