DevPod Conventions-Aware Aider Config

An Aider configuration optimized for DevPod development that automatically reads project conventions and uses the Moonshot Kimi K2 model via Groq for fast, context-aware code generation.

What This Does

This configuration sets up Aider to:

Automatically read `CONVENTIONS.md` before making changes

Use the Moonshot Kimi K2 Instruct model for primary coding tasks

Use a lighter GPT model for simpler operations

Enable caching and diff display for efficient iteration

Disable vim mode and URL detection for cleaner workflow

Configuration

Create a `.aider.conf.yaml` file in your project root:

```yaml

read:

- CONVENTIONS.md

vim: false

detect-urls: false

model: groq/moonshotai/kimi-k2-instruct-0905

timeout: 30

cache-prompts: true

show-diffs: true

weak-model: groq/openai/gpt-oss-20b

```

Model Options

The configuration includes commented alternatives you can swap in:

**Primary model:** `groq/moonshotai/kimi-k2-instruct-0905` (active)

**Alternatives:**

- `codestral/codestral-latest` - Mistral's specialized code model

- `groq/openai/gpt-oss-120b` - Larger open-source GPT variant

- `openrouter/moonshotai/kimi-k2-thinking` - Reasoning-focused Kimi variant

Key Settings

| Setting | Value | Purpose |

|---------|-------|---------|

| `read` | `CONVENTIONS.md` | Auto-load project conventions into context |

| `vim` | `false` | Use standard editing mode |

| `detect-urls` | `false` | Prevent URL auto-detection interference |

| `timeout` | `30` | 30-second API timeout |

| `cache-prompts` | `true` | Speed up repeated operations |

| `show-diffs` | `true` | Display changes before applying |

| `weak-model` | `groq/openai/gpt-oss-20b` | Lighter model for simple tasks |

Usage

1. Create a `CONVENTIONS.md` file in your project documenting coding standards, architecture patterns, and project-specific conventions

2. Add the `.aider.conf.yaml` file to your project root

3. Run `aider` - it will automatically load conventions and use the configured models

4. All edits will respect the conventions defined in your documentation

Why This Works

**Convention-first:** Reading `CONVENTIONS.md` ensures all changes align with project standards

**Fast model:** Groq's inference is optimized for low latency

**Kimi K2:** Strong at following instructions and understanding context

**Diff preview:** See exactly what will change before accepting

**Prompt caching:** Reuses context across requests to reduce costs and latency

Notes

Requires Groq API access for the Moonshot models

The `CONVENTIONS.md` file should be maintained as your project evolves

Swap models by uncommenting alternatives in the config

Timeout can be adjusted for larger codebases or slower connections

DevPod Conventions-Aware Aider Config

DevPod Conventions-Aware Aider Config

What This Does

Configuration

Model Options

Key Settings

Usage

Why This Works

Notes

Reviews (0)