Help — PromptNotes.ai

What is Git?

Git tracks every change ever made to your files — a time machine for your prompts. See who changed what, when, and why, and go back to any version.

📁

Repository

A folder of prompts with complete history

💾

Commit

A saved snapshot with a description

🌿

Branch

A parallel version to experiment safely

🍴

Fork

Your own copy of someone else's repo

⬆️

Push

Upload local commits to PromptNotes.ai

⬇️

Clone

Download a repository locally

💡

No Git needed. You can create and edit prompts entirely through the web interface. Git is for power users who want offline access and automation.

Repository Structure

A prompt repository is a folder of Markdown files. Organize by category:

Directory layout

my-prompt-library/
  ├── README.md
  ├── coding/
  │     ├── python-review.md
  │     └── debugging.md
  ├── writing/
  │     └── blog-post.md
  └── research/
        └── paper-summary.md

Example prompt file

coding/python-review.md

# Python Code Review

**Model:** GPT-4, Claude 3.5
**Task:** Code review

## Prompt

You are an expert Python developer. Review the following
code for correctness, PEP 8 compliance, performance,
and security vulnerabilities.

## Notes
Works best with functions under 100 lines.

Create an Account

1

Click Register

Top-right corner of promptnotes.ai.

2

Fill in your details

Choose a username, email, and a password of at least 12 characters.

3

Add an SSH key (optional)

Settings → SSH Keys → Add Key. Enables passwordless git push.

Create a Repository

1

Click + → New Repository

Top navbar, plus button.

2

Name your collection

Example: gpt4-writing-prompts or claude-coding-assistant.

3

Set visibility

Public repos appear in search and trending. Private repos are invite-only.

4

Initialize with README

Check Initialize this repository to create the first commit automatically.

Upload Prompts via Web

1

Open your repository

promptnotes.ai/username/repo-name

2

New File

Click + → New File. Name it my-prompt.md and paste your content.

3

Commit

Write a message like Add GPT-4 writing prompt and click Commit Changes.

Explore & Search

The search bar on the homepage searches all public repositories. Filter by type:

Repositories — collections by name or description
Files — inside prompt file contents
Commits — by commit message
Branches — branch names
Issues — discussions and improvement requests

Full results open at search.promptnotes.ai/results.

Setup Git

Step 1 — Install Git

# macOS
brew install git

# Ubuntu / Debian
sudo apt install git

Step 2 — Configure identity

git config --global user.name "Your Name"
git config --global user.email "[email protected]"

Step 3 — Generate SSH key

ssh-keygen -t ed25519 -C "[email protected]"
# Press Enter 3 times to accept defaults (no passphrase)

Step 4 — Copy your public key

cat ~/.ssh/id_ed25519.pub
# Output looks like:
# ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAA... [email protected]
# Copy the entire line

🔑

Step 5 — Add key to PromptNotes.ai: Go to promptnotes.ai/user/settings/keys → Add Key → paste the output from Step 4 → click Add Key. Ignore the ssh-keygen verify command Gitea shows — it is optional.

Step 6 — Test SSH connection

ssh -p 2222 -T [email protected]
# Expected response:
# Hi there, promptadmin! You've successfully authenticated...

💡

Use git.promptnotes.ai for SSH — not promptnotes.ai. The main domain routes through Cloudflare which blocks SSH. The git. subdomain points directly to the VPS and allows port 2222.

Basic Git Commands

Clone a repository

git clone ssh://[email protected]:2222/username/repo-name.git
cd repo-name

Add, commit, push

git add my-prompt.md        # stage a file
git add .                    # stage all changes
git commit -m "Add GPT-4 writing prompt"
git push origin main

Branching

git checkout -b claude-variants  # create + switch
git add .
git commit -m "Add Claude-specific variants"
git push origin claude-variants

Recommended Workflow

1

Clone

Terminal

git clone ssh://[email protected]:2222/promptadmin/awesome-chatgpt-prompts.git

2

Branch

Terminal

git checkout -b add-coding-prompts

3

Add, commit, push

Terminal

git add coding/my-prompt.md
git commit -m "Add Python debugging prompt"
git push origin add-coding-prompts

4

Open a Pull Request

Go to your repo → Pull Requests → New Pull Request to propose your additions.

Python Interface

PromptNotes.ai exposes a full REST API. Read, write, and search prompts programmatically — perfect for AI pipelines and automation.

🔑

Generate a token: Settings → Applications → Generate Token. Check all repository scopes. promptnotes.ai/user/settings/applications

Python — setup

pip install requests

GITEA_URL  = "https://promptnotes.ai"
REPO_OWNER = "your-username"
REPO_NAME  = "your-prompt-library"
API_TOKEN  = "your-token"

HEADERS = {
    "Authorization": f"token {API_TOKEN}",
    "Content-Type":  "application/json",
    "User-Agent":    "promptnotes-client/1.0"
}

Reading Prompts

Python — list files

import requests, base64

def list_prompts(owner, repo, path=""):
    url = f"{GITEA_URL}/api/v1/repos/{owner}/{repo}/contents/{path}"
    return requests.get(url, headers=HEADERS).json()

files = list_prompts("promptadmin", "awesome-chatgpt-prompts")
for f in files:
    print(f["name"], f["type"])

Python — read file content

def read_prompt(owner, repo, filepath):
    url  = f"{GITEA_URL}/api/v1/repos/{owner}/{repo}/contents/{filepath}"
    data = requests.get(url, headers=HEADERS).json()
    return base64.b64decode(data["content"]).decode("utf-8")

text = read_prompt("promptadmin", "awesome-chatgpt-prompts", "prompts/coding/python-review.md")
print(text)

Publishing Prompts

Python — create or update a file

import requests, base64

def publish_prompt(owner, repo, filepath, content, message="Update prompt"):
    url      = f"{GITEA_URL}/api/v1/repos/{owner}/{repo}/contents/{filepath}"
    existing = requests.get(url, headers=HEADERS)
    sha      = existing.json().get("sha") if existing.status_code == 200 else None

    payload = {
        "message": message,
        "content": base64.b64encode(content.encode()).decode(),
        "branch":  "main"
    }
    if sha:
        payload["sha"] = sha  # required for updates

    method = "PUT" if sha else "POST"
    r = requests.request(method, url, headers=HEADERS, json=payload)
    return r.status_code

publish_prompt("your-username", "my-prompts",
               "research/summarise-paper.md",
               "# Summarise Research Paper\n...",
               "Add research paper summariser")

Search Integration

Python — search prompts

SEARCH_URL = "https://search.promptnotes.ai"

def search_prompts(query, search_type="repositories", limit=10):
    r = requests.get(f"{SEARCH_URL}/search", params={
        "q": query, "type": search_type, "limit": limit
    })
    data = r.json()
    return data["hits"], data["totalHits"]

hits, total = search_prompts("python debugging")
print(f"Found {total} repositories")
for h in hits:
    print(f"  {h['full_name']} — {h.get('description','')}")

Full pipeline — search, read, use with OpenAI

Python — end-to-end example

import openai

def find_and_run(keyword, user_input):
    # 1. Find a matching prompt on PromptNotes.ai
    hits, _ = search_prompts(keyword, search_type="files", limit=1)
    if not hits:
        raise ValueError(f"No prompt found for: {keyword}")

    owner, repo = hits[0]["repo_full_name"].split("/")

    # 2. Read the prompt from PromptNotes.ai
    prompt_text = read_prompt(owner, repo, hits[0]["path"])

    # 3. Fill in the template placeholder
    filled = prompt_text.replace("{input}", user_input)

    # 4. Call OpenAI
    client = openai.OpenAI()
    resp   = client.chat.completions.create(
        model    = "gpt-4",
        messages = [{"role": "user", "content": filled}]
    )
    return resp.choices[0].message.content

result = find_and_run("python code review", "def add(a,b): return a+b")
print(result)

✅

The motto in action: prompts live on PromptNotes.ai, thrive through community contributions, and reach the world through this API — powering AI applications everywhere.