Error: Devstral 2 (Cloud - Full) - 500

[apaaaa]

Error Details

Model: Devstral 2 (Cloud - Full)
Provider: ollama
Status Code: 500

Error Output

"Internal Server Error"

Additional Context
Please add any additional context about the error here

anytime I set up .continue/rules file I get 500 Error

.continue/rules

rules:

• name: "Full Codebase Access"
  description: "Grant full access to the IntelliJ project for analysis and assistance."
  globs:
  • "**/*.java" # Java files
  • "**/*.kt" # Kotlin files
  • "**/*.xml" # XML files (e.g., Spring Boot configs)
  • "**/*.yml" # YAML files (e.g., application.yml)
  • "**/*.yaml" # YAML files (alternative extension)
  • "**/*.properties" # Properties files
  • "**/*.gradle" # Gradle build files
  • "**/*.gradle.kts" # Gradle Kotlin DSL files
  • "**/pom.xml" # Maven build files
  • "**/build.gradle" # Gradle build files
  • "**/settings.gradle" # Gradle settings files
  • "**/Dockerfile" # Docker configurations
  • "**/*.sql" # SQL files
  • "**/*.md" # Markdown documentation
  • "/src/" # Entire source directory
  • "/test/" # Test directory
  • "/resources/" # Resources directory
    alwaysApply: true

[sestinj]

Potentially related issues:

• Failure to parse "rules" block with custom config and using CLI #9182 - Failure to parse "rules" block with custom config
• More helpful warnings for invalid .continue/rules #9548 - More helpful warnings for invalid .continue/rules
• Some rules are treated as prompts #8885 - Some rules are treated as prompts

[apaaaa]

that is my config.yaml
maybe the issue is that here and in the rules file also rules: section?

name: Local Config
version: 1.0.0
schema: v1

rules:
  - name: Spring Boot Expert
    rule: "You are an expert Spring Boot and AWS architect. Focus on: Clean architecture (hexagonal, layered), Spring Boot 3.x best practices, AWS serverless patterns, Microservices design, RAG system architecture. When you need the latest version numbers or current documentation that might have changed recently, use web search. Always explain trade-offs and provide runnable code examples."
    alwaysApply: true

models:
  - name: Devstral 2 (Cloud - Full)
    provider: ollama
    model: devstral-2:123b-cloud
    apiBase: https://api.ollama.com
    apiKey: 
    capabilities:
      - tool_use
    roles:
      - chat
      - edit
      - apply
      - autocomplete

  - name: Devstral 2 Small (Cloud - Vision)
    provider: ollama
    model: devstral-small-2:24b-cloud
    apiBase: https://api.ollama.com
    apiKey: 
    capabilities:
      - tool_use
    roles:
      - chat
      - edit
      - apply

  - name: Qwen3 Coder 480B (Cloud - Backup Heavy)
    provider: ollama
    model: qwen3-coder:480b-cloud
    apiBase: https://api.ollama.com
    apiKey: 
    capabilities:
      - tool_use
    roles:
      - chat
      - edit
      - apply
      - autocomplete

  - name: Qwen3-VL (Cloud - Screenshots)
    provider: ollama
    model: qwen3-vl:235b-cloud
    apiBase: https://api.ollama.com
    apiKey: 
    capabilities:
      - tool_use
    roles:
      - chat

experimental:
  webSearch:
    provider: brave
    apiKey: ""

[RomneyDa]

@apaaaa have you had success trying any other models? Wondering if it's an issue on the ollama side

[gdellis]

I just use https://olama.com/v1 as my api base for ollama cloud models. I seems to work well. Although occasionally it has issues with tool use. You can refer to ollama's docs on the different integrations, such as codex here https://docs.ollama.com/integrations/codex

[gdellis]

Here is my yaml anchor for ollama cloud that is working for me:

.ollama_cloud: &ollama_cloud
  provider: openai
  apiBase: https://ollama.com/v1
  apiKey: ${{ secrets.OLLAMA_API_KEY }}
  capabilities: []

models:
  - name: devstral-2:123b
    model: devstral-2:123b
    <<: *ollama_cloud

NOTE: One thing to note is that if you use the cloud models with https://ollama.com/v1 , is you don't use the cloud suffix in the model name

You can also use the .env file ( ~/.continue/.env) to set your OLLAMA_API_KEY, which is referenced in my config above. When continue starts it reads your .env file for secrets and other environmental stuff

EXAMPLE:

OLLAMA_API_KEY=123456789.... # Replace with your actual key

[apaaaa]

Hi @gdellis and @RomneyDa
now it is working musch better, I tailored my settings based on yours. Thanks.

My environment:

• IntelliJ
• continue.dev
• Ollama cloud subscribtion

I get this error when I ask the coding assistant in IntelliJ (continue.dev) to fix some code. Actually it did, but still throwing these errors.

can you pls check my configs maybe you spot the issue

config.yaml:
name: Local Config
version: 1.0.0
schema: v1

models:
  - name: Devstral 2 (Cloud - Full)
    provider: openai
    model: devstral-2:123b
    apiBase: https://ollama.com/v1
    apiKey: 
    capabilities:
      - tool_use
    roles:
      - chat
      - edit
      - apply

  - name: Devstral 2 Small (Cloud - Vision)
    provider: openai
    model: devstral-small-2:24b
    apiBase: https://ollama.com/v1
    apiKey: 
    capabilities:
      - tool_use
    roles:
      - chat
      - edit
      - apply

  - name: Qwen3 Coder 480B (Cloud - Backup Heavy)
    provider: openai
    model: qwen3-coder:480b
    apiBase: https://ollama.com/v1
    apiKey: 
    capabilities:
      - tool_use
    roles:
      - chat
      - edit
      - apply
      - autocomplete

  - name: Qwen3-VL (Cloud - Screenshots)
    provider: openai
    model: qwen3-vl:235b
    apiBase: https://ollama.com/v1
    apiKey: 
    capabilities:
      - tool_use
    roles:
      - chat

experimental:
  webSearch:
    provider: brave
    apiKey: 

.continue/rules

rules:
  - name: Spring Boot Expert
    rule: "You are an expert Spring Boot and AWS architect. Focus on: Clean architecture (hexagonal, layered), Spring Boot 3.x best practices, AWS serverless patterns, Microservices design, RAG system architecture. When you need the latest version numbers or current documentation that might have changed recently, use web search. Always explain trade-offs and provide runnable code examples."
    alwaysApply: true

  - name: All Project Files
    rule: "Analyze all project files including documentation, configuration, build scripts, and CI/CD workflows. Apply appropriate best practices for each file type."
    globs:
      - "**/*"

  - name: Documentation Files
    rule: "When reviewing documentation files (.md, .txt, .adoc), focus on clarity, completeness, and accuracy. Ensure technical documentation aligns with actual implementation."
    globs:
      - "**/*.md"
      - "**/*.txt"
      - "**/*.adoc"

  - name: Configuration Files
    rule: "When reviewing configuration files, check for security best practices, environment-specific settings, and proper organization. Pay attention to sensitive data handling."
    globs:
      - "**/*.properties"
      - "**/*.yml"
      - "**/*.yaml"
      - "**/*.json"
      - "**/*.toml"
      - "**/*.env*"

  - name: Build and Deployment Files
    rule: "When reviewing build and deployment files, ensure they follow infrastructure-as-code principles, security best practices, and maintainability standards."
    globs:
      - "**/Dockerfile*"
      - "**/docker-compose*"
      - "**/pom.xml"
      - "**/build.gradle*"
      - "**/.github/workflows/*"
      - "**/.gitlab-ci.yml"
      - "**/Jenkinsfile*"

  - name: Java Code Analysis
    rule: "When analyzing Java code, focus on Spring Boot best practices, clean architecture, proper error handling, security considerations, and performance optimizations."
    globs:
      - "**/*.java"

  - name: Static Web Assets
    rule: "When reviewing static web assets (HTML, CSS, JS), ensure they follow modern frontend best practices, accessibility standards, and are properly integrated with Spring Boot backend APIs. Check for security vulnerabilities like XSS and proper CSRF protection."
    globs:
      - "**/src/main/resources/static/**/*"
      - "**/src/main/resources/templates/**/*"
      - "**/*.html"
      - "**/*.css"
      - "**/*.js"`