grindx

Distraction-free DSA practice in your terminal.

Why grindx?

• Local-first — problem browsing, progress, and solutions stay local; AI review and testcase bundle downloads are optional
• Terminal-native — practice DSA without leaving your terminal or opening a browser
• Multiple sheets — Striver A2Z (316), Blind 75, NeetCode 150, Grind 75 built-in
• 5 languages — Python, Go, C++, Java, JavaScript — switch on the fly with Ctrl+L
• Progress tracking — solved/in-progress/not-started states, streaks, best times
• Bookmarks & filters — filter by difficulty (Easy/Medium/Hard) or bookmarked problems

Install

pip install grindx

Or with pipx (recommended for CLI tools):

pipx install grindx

Or run from source:

git clone https://github.com/xghostient/grindx.git
cd grindx
python3 -m venv .venv && source .venv/bin/activate
pip install -e .
grindx

Local testcase bundle

PyPI installs keep the package lean and do not ship the full bundled testcase corpus. To enable local Ctrl+R test runs, fetch the external testcase bundle once:

grindx --fetch-testcases

By default this pulls from:

https://github.com/grindxhq/dsa-catalog/releases/latest/download/manifest.json

Override the manifest source if needed:

grindx --fetch-testcases --testcase-manifest-url https://example.com/manifest.json
# or
export GRINDX_TESTCASE_MANIFEST_URL=https://example.com/manifest.json

Usage

grindx            # CLI entry point
python -m grindx  # or as a module

Navigation

Filters

Solve Screen

Features

Split-pane editor — problem description on the left, code editor with syntax highlighting on the right.

Auto-timer — starts when you open a problem, tracks your best solve time.

Three-state tracking — each problem shows as not started (○), in progress (◐), or solved (✓).

Stats dashboard — overall progress, per-difficulty breakdown, per-topic progress bars, current streak, and top 10 best times.

Sheet-agnostic — built-in sheets live inside the package (grindx/sheets/). Format:

{
  "Topic Name": ["problem-name-1", "problem-name-2"],
  "Another Topic": ["problem-name-3"]
}

Progress safety — automatic backups with corruption recovery. Progress, solutions, backups, and downloaded testcase bundles are stored in ~/.grindx/ so they persist across installs and upgrades.

AI Review (optional)

Press Ctrl+E on the solve screen to get AI-powered feedback on your solution — correctness, edge cases, complexity analysis, and a pass/fail verdict.

Setup

Set two environment variables:

export GRINDX_AI_PROVIDER=groq              # or ollama, anthropic, openai
export GRINDX_AI_MODEL=llama-3.3-70b-versatile  # optional, sensible defaults per provider
export GRINDX_AI_KEY=gsk_...                # not needed for ollama
export GRINDX_AI_URL=https://custom.api/v1  # optional, auto-detected per provider

Or create ~/.grindx.toml:

[ai]
provider = "groq"
model = "llama-3.3-70b-versatile"
api_key = "gsk_..."

Supported Providers

Any OpenAI-compatible API works — set provider to openai and add base_url.

Custom Sheets & Problems

You can add your own problem sheets and problems alongside the built-in ones.

Custom sheet

Create a JSON file in ~/.grindx/sheets/:

// ~/.grindx/sheets/my-list.json
{
  "Arrays": ["two-sum", "best-time-to-buy-and-sell-stock", "my-custom-problem"],
  "Strings": ["valid-anagram", "another-custom-problem"]
}

You can mix built-in problem IDs (like two-sum) with your own custom problem IDs. Run grindx --list-problems to see all available built-in problem IDs.

Custom problems

Create a folder per problem under ~/.grindx/problems/:

// ~/.grindx/problems/my-custom-problem/problem.json
{
  "id": "my-custom-problem",
  "name": "My Custom Problem",
  "difficulty": "Medium",
  "category": "Arrays",
  "description": "Given an array of integers, find ...",
  "examples": [
    {"input": "nums = [1, 2, 3]", "output": "6"},
    {"input": "nums = []", "output": "0"}
  ],
  "constraints": "1 <= nums.length <= 10^4",
  "python_template": "def solve(nums):\n    pass\n",
  "go_template": "func solve(nums []int) int {\n\n}\n",
  "cpp_template": "int solve(vector<int>& nums) {\n\n}\n",
  "java_template": "class Solution {\n    public int solve(int[] nums) {\n\n    }\n}\n",
  "js_template": "function solve(nums) {\n\n}\n"
}

Optional local judge assets for a custom problem:

~/.grindx/problems/my-custom-problem/testcases.json
~/.grindx/problems/my-custom-problem/judges/python.py
~/.grindx/problems/my-custom-problem/judges/go.go
~/.grindx/problems/my-custom-problem/judges/cpp.cpp
~/.grindx/problems/my-custom-problem/judges/java.java
~/.grindx/problems/my-custom-problem/judges/javascript.js

You only need to include the template keys and local judge assets for the languages you care about.

Legacy ~/.grindx/problems/custom.json is still loaded for compatibility, but new custom problems should use the folder-based format above.

Custom sheets show up on the dashboard alongside the built-in ones.

Built with

• Textual — TUI framework
• tree-sitter — syntax highlighting

License

MIT