PromptHive 🐝

Lightning-fast open source prompt manager for developers. Terminal-native, sub-15ms operations, works with any AI tool. 1000+ organic downloads.

🚀 Website: prompthive.sh | 🤝 Companion Tool: CalmHive for background AI processing

TL;DR

# Install (Latest: v0.2.8)
cargo install prompthive
# OR: curl -sSL https://prompthive.sh/install.sh | bash

# Multiple Workflows - Choose Your Style

# 1. Perfect commit messages (PRODUCTION READY ✅)
git diff --staged | ph use essentials/commit              # See prompt output
claude "$(git diff --staged | ph use essentials/commit)"  # Interactive AI

# 2. Instant debugging (BATTLE TESTED ✅)
cat error.log | ph use essentials/debug | llm -p > analysis.md    # Non-interactive
claude "$(cat error.log | ph use essentials/debug)"              # Interactive debug session

# 3. Complete PR workflows
git diff main...HEAD | ph use essentials/review | claude -p | gh pr create --body-file -  # Full automation

Why? Because copy-pasting from ChatGPT history takes 30+ seconds. PromptHive operations average 8ms.

Current Version: 0.2.8 - Open source with community registry, instant authentication, team collaboration, and prompt sharing. All features included, no paid tiers.

🚀 Philosophy: Terminal-First Development | Spec-Driven AI Engineering 📁 Complete Guide: prompthive.sh | 🤖 Background AI: calmhive.com

Core Promise: Every command under 15ms. It's a TOOL, not a LIBRARY.

Unix Philosophy: We don't run AI. We manage prompts perfectly, and pipe them to ANY AI tool you prefer. Like npm doesn't run JavaScript, we don't run prompts. We just make them instantly accessible.

# PromptHive + Your Favorite AI Tools - Three Patterns:

# 1. Command Substitution (Interactive AI):
claude "$(git diff | ph use essentials/review)"          # Interactive session
llm "$(ph use essentials/debug)" error.log              # Chat with context

# 2. Piped with -p flag (Non-interactive/Scripting):
git diff --staged | ph use essentials/commit | claude -p | git commit -F -
cat api.py | ph use essentials/review | llm -p > api-docs.md

# 3. Direct Output (Auto-clipboard):
ph use essentials/debug      # Output + clipboard for pasting into ChatGPT UI
ph use essentials/commit -q  # Quiet mode (no clipboard)

🚀 Installation

Cargo (Rust) - Available Now ✅

# Install from crates.io
cargo install prompthive

# Verify installation
ph --version

Quick Install Script

# Automatic installation with platform detection
curl -sSL https://prompthive.sh/install.sh | bash

# Verify installation
ph --version

Current Status: Version 0.2.8 is available on crates.io. The install script automatically downloads pre-built binaries for your platform.

Shell Completions

After installation, set up shell completions for the best experience:

Bash

ph completion bash > ~/.bash_completion.d/prompthive
source ~/.bash_completion.d/prompthive

Zsh

ph completion zsh > ~/.zsh_completions/_prompthive
# Add to ~/.zshrc: fpath=(~/.zsh_completions $fpath)

Fish

ph completion fish > ~/.config/fish/completions/ph.fish

Quickstart

# Use built-in essentials (ready to go!)
git diff --staged | ph use essentials/commit | llm   # Perfect commit messages
ph use essentials/debug "Fix auth timeout error" | claude # Analyze any error

# Use immediately with real workflows
git diff main...HEAD | ph use essentials/review | llm     # Generate PR descriptions
cat main.py | ph use essentials/review | llm    # Review any code

# Use with ANY AI tool (copies to clipboard if terminal, pipes if piped)
ph use essentials/debug "Fix auth timeout" | llm     # LLM by Simon Willison
git diff --staged | ph use essentials/commit | aichat    # AIChat
git diff main...HEAD | ph use essentials/review | claude  # Interactive PR description
cat legacy-code.js | ph use 10x/refactor | mods    # Mods
cat api.py | ph use essentials/review | sgpt      # Shell GPT

# Lightning-fast operations
ph ls                                      # List all prompts
ph f commit                                # Fuzzy search essentials
ph u essentials/debug-error                # Short aliases work

# Share and organize with banks
ph bank publish my-workflow-prompts
# Create custom banks from your favorite prompts
mkdir -p ~/.prompthive/banks/team
cp ~/.prompthive/prompts/*.md ~/.prompthive/banks/team/

"But Why Not Just..."

1000+ developers chose PromptHive. Here's why existing alternatives don't scale:

...use text files in ~/prompts/?

Reality check: You tried this. It works until you have 20+ prompts, then:

Can't fuzzy search across them in 8ms (try find + grep—it's slow)
No smart matching (ph u au → auth-basic instantly)
No metadata, descriptions, or organization
No team sharing or version control
Biggest issue: You stop using them because access isn't instant

...use shell aliases?

# This worked for 5 prompts. What about 50?
alias debug="echo 'Debug this error and suggest a fix'"
alias commit="echo 'Generate a commit message from the diff'"
alias review="echo 'Review this code for issues'"
alias api="echo 'Design a REST API for...'"
# Your .bashrc becomes unmaintainable

# vs PromptHive:
ph f debug   # Fuzzy finds essentials/debug-error in 8ms
ph f api     # Finds all API-related prompts instantly

...use Neovim with snippets?

For developers who live in Neovim: You could build this with:

Snippet plugins (20 minutes setup)
Custom fuzzy search (another plugin)
Template variables (more configuration)
Team sharing (Git repos + more setup)

Reality: Most developers want prompts to "just work" across all contexts—terminal, browser, any AI tool. PromptHive is purpose-built for this.

...use ChatGPT/Claude history search?

30+ seconds to find that perfect prompt from last week
Lost when you clear browser history
Can't compose: No cat error.log | [your-prompt] | claude
No iteration: Can't improve prompts over time
No sharing: Your team reinvents the same prompts

...use GitHub Gists or company wikis?

Network dependency: Fails when WiFi is down
No terminal integration: Can't pipe directly
Context switching: Breaks flow between terminal and browser
Slow: 2-5 seconds minimum to access

The Real Differentiator: Workflow Velocity

When prompt access drops below 80ms, something profound happens: You stop thinking about the tool and start thinking with it.

This isn't about saving 30 seconds once. It's about saving 5 minutes, 20 times per day, while maintaining flow state.

Commands (v0.2.8)

ph use <name>         # Use a prompt (u) - auto-clipboard, save, append, file
ph show <name>        # Display prompt (s) - with I/O options
ph new <name>         # Create prompt (n) - smart detection
ph edit <name>        # Edit prompt (e) - opens $EDITOR
ph find <query>       # Search prompts (f) - fuzzy matching
ph ls                 # List prompts (l) - see all
ph delete <name>      # Delete prompt (d) - with confirm
ph clean <text>       # Clean and format text with AI-style processing

# Just run 'ph' to launch the TUI (like lazygit)
ph                    # Launch interactive TUI

### Advanced Features
```bash
ph compose <prompts>  # Chain prompts together
ph stats              # Usage analytics dashboard
ph completion <shell> # Generate shell completions
ph login              # Authenticate with registry for sync
ph sync               # Sync prompts with cloud ✅ WORKING
ph sync status         # Check sync status and conflicts ✅ WORKING
ph sync push          # Push local changes to cloud ✅ WORKING
ph sync pull          # Pull cloud changes locally ✅ WORKING

🚀 Power User Examples

Built-in Template Variables

# Prompts automatically support built-in variables
ph new standup "Daily standup for {date} by {user} on {hostname}"
ph new bug-report "Bug in {pwd} on branch {git_branch} ({git_status})"

# Environment variables work too
ph new deploy "Deploy to {env:AWS_REGION} in {env:ENVIRONMENT}"

# Variables are replaced when you use the prompt
ph use standup | llm
# Output includes current date, username, and hostname automatically

Version Control - Never Lose a Good Prompt

# Tag important versions of your prompts
ph version api-design v1.0 -m "Initial stable API prompt"
ph version api-design v1.1 -m "Added error handling section"
ph version api-design v2.0 -m "Complete rewrite for REST best practices"

# View history
ph versions api-design
# 📚 Version history for 'api-design'
# 📌 v2.0 (a3b4c5d6) - Complete rewrite for REST best practices
# 📌 v1.1 (87654321) - Added error handling section
# 📌 v1.0 (12345678) - Initial stable API prompt

# Rollback when needed
ph rollback api-design v1.1
# ✅ Rolled back 'api-design' to version 'v1.1'

Diff & Merge - Collaborate on Prompts

# Compare prompt versions
ph diff api-v1 api-v2
ph diff api-v1 api-v2 --format side-by-side
ph diff local-prompt team/shared-prompt --context 10

# Merge improvements
ph merge team/api-enhanced api-design --backup
ph merge experimental-fix stable-prompt --preview

Web Dashboard - Visual Analytics

# Open interactive dashboard
ph web                    # Full dashboard
ph web stats             # Usage analytics
ph web prompts           # Browse all prompts
ph web --no-browser      # Generate HTML only

🎯 Unified I/O Design - Smart Defaults

PromptHive uses intelligent defaults based on context, reducing typing while maintaining flexibility:

Text Transformation Commands (use, show, clean)

In Terminal (TTY):

Auto-copies to clipboard by default
No stdout output (content is on clipboard)
Use -q to suppress clipboard and output to stdout instead

When Piping:

Outputs to stdout for pipe compatibility
No auto-clipboard (use -c to force clipboard)

# Terminal usage - auto-clipboard
ph clean "messy text"                    # ✓ Copied to clipboard (2ms)
ph use api-design "Create user endpoint" # ✓ Copied to clipboard (3ms)

# Piping - outputs to stdout
echo "messy text" | ph clean | ph use formatter | claude -p

# Force clipboard even when piping
echo "important" | ph clean -c          # Copies AND outputs

# Quiet mode in terminal
ph clean "text" -q                      # Outputs to stdout, no clipboard

Query Commands (ls, find)

Always outputs to stdout (auto-clipboard would be too noisy)

Use -c to explicitly copy results
Use -s to save as a new prompt

ph ls                                    # Lists to stdout
ph find "api" -c                        # Find and copy results
ph ls -s "my-prompt-list"               # Save list as a new prompt

Universal I/O Flags

Every command supports these consistent flags:

-s NAME - Save output as a new prompt
-a NAME - Append output to existing prompt
-c - Force copy to clipboard
-f PATH - Write to file (with bidirectional sync)
-q - Quiet mode (suppress default behaviors)
-e - Edit before output (where applicable)

Smart File Operations

The -f flag creates bidirectional sync by default:

# Create prompt with automatic file sync
ph new "API guidelines" -f              # Creates ./api-guidelines.md
ph new "API guidelines" -f api.md       # Creates ./api.md

# Smart naming when path not provided
ph use api-design "endpoints" -f        # Creates ./api-design-output.md

# Bidirectional sync means:
# - Edit the file → prompt updates
# - Edit the prompt → file updates
# - Always stay in sync!

Composable Flags

All flags work together sensibly:

# Clean, save, append, copy, and write to file
ph clean "text" -s cleaned -a log -c -f output.md

# Use prompt, edit result, save, and sync to file
ph use template -e -s edited -f template-output.md

# Find prompts, save results, copy to clipboard
ph find "api" -s search-results -c

Design Philosophy

Smart defaults: Do the right thing based on context
Explicit overrides: Flags always override defaults
Composability: All flags combine logically
No surprises: Predictable behavior in all contexts

Configuration

Disable Logging

# Temporary (this session only)
export PROMPTHIVE_LOG_LEVEL=error    # Only errors
export PROMPTHIVE_LOG_LEVEL=off      # Complete silence

# Permanent (add to ~/.zshrc or ~/.bashrc)
echo 'export PROMPTHIVE_LOG_LEVEL=error' >> ~/.zshrc
source ~/.zshrc

Other Environment Variables

PROMPTHIVE_BASE_DIR=/custom/path     # Change storage location
PROMPTHIVE_EDITOR=code               # Set preferred editor
PROMPTHIVE_LOG_FORMAT=json           # JSON logs for production

Creative Usage Patterns

🎯 Daily Journaling & Reflection

# Create a journal bank
ph new journal/daily "Date: {date}\n\nToday I learned: "
ph new journal/gratitude "Three things I'm grateful for:\n1. "
ph new journal/standup "Yesterday: \nToday: \nBlockers: "

# Daily workflow
claude "$(date | ph use journal/daily)"
ph use journal/gratitude -s entries/$(date +%Y-%m-%d)

🧠 Learning & Study Assistant

# Create study prompts
ph new study/explain "Explain {input} like I'm 5"
ph new study/test "Create 5 quiz questions about: {input}"
ph new study/summarize "Key points from this text: {input}"

# Study session
cat lecture-notes.md | ph use study/summarize | claude -p > summary.md
claude "$(cat chapter-3.txt | ph use study/test)"

💼 Meeting & Communication Templates

# Professional templates
ph new work/email-followup "Subject: Follow-up from our {date} meeting\n\nHi {name},\n\nThank you for..."
ph new work/1-on-1 "1-on-1 with {manager}\n\nAgenda:\n- Career development\n- Current projects\n- Feedback\n\nNotes: "
ph new work/proposal "Proposal: {title}\n\nProblem: \nSolution: \nImpact: \nTimeline: "

# Quick usage
ph use work/email-followup "meeting yesterday" | pbcopy
claude "$(ph use work/proposal 'Implement CI/CD Pipeline')"

🎨 Creative Writing & Content

# Writing helpers
ph new write/character "Create a character profile:\nName: {input}\nTraits: "
ph new write/plot-twist "Given this plot: {input}\n\nSuggest 3 unexpected twists:"
ph new write/blog-outline "Blog post about: {input}\n\nOutline with sections:"

# Content creation
claude "$(ph use write/blog-outline 'Terminal productivity')"
echo "Detective Jane Smith" | ph use write/character | llm -p

🔧 DevOps & Operations

# Runbook templates
ph new ops/incident "Incident: {title}\nSeverity: \nImpact: \nMitigation: "
ph new ops/postmortem "Postmortem for: {incident}\n\nWhat happened: \nRoot cause: \nLessons learned: "
ph new ops/deploy-checklist "Deploy {service} to {env}\n\n[ ] Tests pass\n[ ] Migrations run\n[ ] Monitoring updated\n[ ] Rollback plan"

# Incident response
ph use ops/incident "Database connection timeout" -s incidents/$(date +%s)
claude "$(kubectl logs -n prod app-pod --tail=100 | ph use ops/debug)"

📊 Data Analysis & Research

# Analysis templates
ph new data/hypothesis "Hypothesis: {input}\n\nTest with: \nExpected outcome: "
ph new data/findings "Analysis of: {dataset}\n\nKey findings: \nLimitations: \nNext steps: "
ph new data/visualize "Data: {input}\n\nSuggest 3 visualization types and why:"

# Research workflow
cat experiment-results.csv | ph use data/findings | claude -p > analysis.md
claude "$(ph use data/hypothesis 'User engagement increases with dark mode')"

🎮 Personal Productivity

# Life management
ph new life/decision "Decision: {input}\n\nPros:\nCons:\nAlternatives:"
ph new life/habit "Track habit: {habit}\n\nDate: {date}\nCompleted: [ ]\nNotes: "
ph new life/goals "Goal: {input}\n\nWhy it matters: \nSuccess criteria: \nFirst step: "

# Daily use
claude "$(ph use life/decision 'Accept job offer at startup')"
ph use life/habit "Morning meditation" -a habits/meditation-log

🚀 API & Integration Patterns

# Combine with other tools
alias morning='claude "$(git log --since=yesterday | ph use journal/standup)"'
alias debug='ph use essentials/debug | tee debug.log | claude'
alias review='gh pr view --json files | ph use essentials/review | claude -p'

# Scheduled prompts
crontab -e
# 0 9 * * 1 ph use work/weekly-goals | mail -s "Weekly Goals" me@example.com
# 0 17 * * * git log --since="9am" | ph use journal/daily | ph save today

Performance

PromptHive operations are optimized for instant prompt access:

$ time ph use essentials/commit
✓ Prompt loaded and copied to clipboard (3ms)
real    0m0.045s

$ ph ls
📋 Your prompts: (12ms)
  essentials/commit       - Semantic commit messages
  essentials/debug-error  - Error analysis & fixes
  essentials/pr          - Pull request descriptions
  essentials/code-review - Code review analysis
  essentials/refactor    - Code refactoring

What these times represent:

PromptHive operations: File I/O, prompt processing, fuzzy matching (~2-50ms)
AI processing time: Separate - depends on your AI tool (5-30+ seconds)
Value proposition: Zero-latency prompt access vs. writing prompts from scratch

The real speed gain: Instead of spending 2-5 minutes crafting the perfect prompt, you get battle-tested prompts instantly.

Partial Matching (Like JJ)

PromptHive uses smart partial matching:

$ ph use d
# If unique, uses the match
# If ambiguous, shows options:
Error: Multiple matches. Did you mean:
  essentials/debug-error    (de)   - Error analysis & fixes
  essentials/document       (do)   - Documentation generation
  custom/deploy            (dep)  - Deployment scripts

$ ph use de  # Now unique - uses 'essentials/debug-error'

The Magic: Composition

PromptHive is the ONLY tool that lets you chain prompts:

# Fix a bug completely
cat error.log | \
  ph use analyze-error | \
  ph use find-root-cause | \
  ph use generate-fix | \
  ph use add-tests | \
  claude

# Each prompt transforms the previous output
# Total time: 250ms for 5 prompts

Real-World Usage

Daily Developer Workflows

# 🔥 Auto-generate perfect commit messages
git diff --staged | ph use essentials/commit | llm

# 🐛 Debug errors instantly
cat error.log | ph use essentials/debug | aichat
ph use essentials/debug "Fix timeout in API calls" | claude

# 📝 Generate PR descriptions
git diff main...HEAD | ph use essentials/review | claude

# 🧪 Create tests from code
cat src/api.js | ph use essentials/review | llm > src/api.test.js

# 📚 Document your code
cat complex-function.py | ph use essentials/docstring | llm

# 🔍 Code review helper
git show HEAD | ph use essentials/review | aichat > review-notes.md

Advanced Workflows

# Chain multiple prompts for complex tasks
cat buggy-code.js | \
  ph use essentials/debug | llm | \
  ph use essentials/refactor | llm | \
  ph use essentials/generate-tests | llm > fixed-code.js

# Morning standup automation
git log --since=yesterday --oneline | \
  ph use essentials/standup | llm > standup.md

# Refactor legacy code
cat old-api.py | \
  ph use essentials/refactor | llm | \
  ph use essentials/add-types | llm > modern-api.py

🧙 Workflow Automation Examples

# AI-Powered Git Workflow
alias smart-commit='git diff --staged | ph use commit-message | llm | git commit -F -'
alias smart-pr='git diff main...HEAD | ph use essentials/review | claude -p | gh pr create --body-file -'

# Automated Code Review Pipeline
function ai-review() {
  git diff main...HEAD | \
    ph compose check-style,find-bugs,suggest-improvements | \
    llm > review-$(date +%Y%m%d).md
}

# Daily Standup Generator
function standup() {
  echo "Yesterday: $(git log --since=yesterday --author=$(git config user.name) --oneline)" | \
    ph use standup --edit | \
    pbcopy
}

# Smart Documentation Generator
function doc-this() {
  cat "$1" | \
    ph compose analyze-code,generate-docs,add-examples | \
    llm > "${1%.*}.docs.md"
}
  ph use modernize-code | llm | \
  ph use add-types | llm > modern-api.py

🔰 Shell Pipes Primer (for Terminal Newbies)

The Basics

# The pipe | sends output from one command to another
echo "hello" | cat           # Sends "hello" to cat

# Redirect > saves output to a file
echo "hello" > file.txt      # Saves "hello" to file.txt

# Append >> adds to a file without overwriting
echo "world" >> file.txt     # Adds "world" to file.txt

# Input < reads from a file
cat < file.txt               # Reads content from file.txt

With PromptHive

# Simple: Copy to clipboard (when in terminal)
ph use api-design           # Just copies, no pipe needed

# Pipe to AI: Send prompt to AI tool
ph use api-design | llm     # Prompt → AI tool

# Chain: Multiple operations
cat code.js | ph use review | llm | ph new review-result

# Save: Capture output
ph use api | llm > response.txt

# Combine: Mix different sources
ph use debug "Fix this: $(cat error.log)" | aichat

When to Use What

No pipe: When you just want to copy a prompt
Single pipe |: When sending to one AI tool
Multiple pipes |: When chaining operations
Redirect >: When saving results
Cat/Echo: When adding context to prompts

🏦 Prompt Banks - Instant Productivity

Built-in Prompts Ready to Use

# essentials/ - Core developer workflows
ph use essentials/commit         # Generate conventional commit message from diff
ph use essentials/debug          # Analyze errors and suggest fixes
ph use essentials/review         # Comprehensive code review

# 10x/ - Advanced productivity workflows
ph use 10x/afk-task             # Long-running autonomous task for calmhive
ph use 10x/fix-tests            # Systematically fix all failing tests
ph use 10x/refactor             # Refactor code for clarity and performance

Create Your Own Banks

# Option 1: From existing prompts
mkdir -p ~/.prompthive/banks/myteam
cp ~/.prompthive/prompts/*.md ~/.prompthive/banks/myteam/

# Option 2: From Claude commands (if you have them)
mkdir -p ~/.prompthive/banks/claude
cp ~/.claude/commands/*.md ~/.prompthive/banks/claude/
# Now use: ph use claude/command-name

# Option 3: Create specialized banks
mkdir -p ~/.prompthive/banks/{backend,frontend,devops}
ph new backend/api-design "Design REST APIs..."
ph new frontend/component "React component template..."
ph new devops/deploy "Deployment checklist..."

Sharing Banks with Your Team

# Share via Git (recommended)
cd ~/.prompthive/banks/myteam
git init && git add .
git commit -m "feat: team prompt collection"
git remote add origin github.com/yourteam/prompts
git push

# Team members clone it
cd ~/.prompthive/banks
git clone github.com/yourteam/prompts team
# Now everyone can: ph use team/api-design

# Keep banks in sync
cd ~/.prompthive/banks/team && git pull

# Or use PromptHive's built-in sharing
ph bank publish myteam        # Publish to registry
ph bank install @user/team    # Install from registry

Real Workflows with Banks

# Developer workflows
git diff --staged | ph use essentials/commit | llm
cat error.log | ph use essentials/debug-error | claude
git diff main | ph use essentials/review | llm > review.md

# Professional communication
ph use professional/email-reply "Thanks for your proposal..." | llm
ph use professional/meeting-notes "Sprint planning discussion" | llm
ph use professional/status-update "Project Alpha Week 3" | claude

# Infrastructure & DevOps
ph use devops/dockerfile "Node.js microservice" | llm > Dockerfile
ph use devops/ci-pipeline "React app with tests" | claude > .github/workflows/ci.yml
ph use devops/kubernetes "web API with Redis" | llm > k8s-manifests.yaml

# Design patterns
ph use coding-patterns/api-design "User management system" | claude
ph use coding-patterns/error-handling "Payment processing" | llm

# Advanced workflows
ph use 10x/afk-task "Build authentication system" | claude
ph use workflow/analyze "Review codebase structure" | llm
ph use claude-commands/workflow-manager "Deploy to production" | claude

# Create aliases for common workflows
alias commit='git diff --staged | ph use essentials/commit | llm'
alias email='ph use professional/email-reply'
alias dockerize='ph use devops/dockerfile'

Building PromptHive with Itself

We built PromptHive using PromptHive:

# Bootstrap phase
ph new rust-cli
ph use rust-cli | cursor

# Every feature after
ph new add-command
ph use add-command | claude

# Test and iterate
ph search "rust testing"
ph use rust-test | cursor

File Structure

Dead simple, just markdown files:

~/.prompthive/
├── prompts/          # Your personal prompts
│   ├── api.md
│   ├── auth.md
│   └── test.md
├── banks/           # Organized prompt collections
│   ├── essentials/  # Built-in essentials
│   ├── 10x/        # Productivity workflows
│   ├── workflow/   # Complex processes
│   └── myteam/     # Your custom banks
├── registry/        # Cached registry packages
└── config.toml      # Configuration

Prompt Format

Just markdown with optional frontmatter:

---
id: api
description: REST API design
---

Design a REST API with these requirements:
- Resource: {resource}
- Operations: {operations}
- Include error handling
- Follow REST best practices

Why PromptHive?

Speed: 8ms operations are addictive (ChatGPT: 30+ seconds)
Simple: Just like ls and cat - no learning curve
Universal: Works with ALL AI tools, not locked to one
Offline: Your prompts work without internet
Composable: Chain prompts like Unix commands
Smart I/O: Auto-clipboard, save, append - works how you think

8ms vs 30 seconds. Every time. That's the difference.

Features

Core Features (Production Ready)

✅ Lightning-fast prompt management - Create, edit, search, and use prompts
✅ Intelligent fuzzy matching - ph u ap → api-design instantly
✅ Universal shell integration - Copy to clipboard or pipe to ANY tool
✅ Smart I/O operations - Auto-clipboard, save, append, file output
✅ Text cleaning & formatting - AI-style text processing with ph clean
✅ Smart completions - Tab completion for bash, zsh, fish
✅ Cross-platform - Works on macOS, Linux, Windows
✅ Prompt composition - Chain prompts for complex workflows
✅ Magic link authentication - Secure login for registry sync
✅ Performance guarantee - All operations under 15ms

Advanced Features (All Included - Free & Open Source)

✅ Registry sync - Cloud backup and device sync
✅ Magic link authentication - Passwordless secure login
✅ Team collaboration - Private prompt banks and sharing
✅ Prompt sharing - Public and invite-based prompt sharing
✅ Community registry - Growing library of quality prompts
✅ Usage analytics - Productivity tracking with achievements
✅ Web dashboard - Visual prompt management and statistics

Development Build

# Development build (all features)
cargo build --release

# Run tests
cargo test --release

# Performance benchmark
cargo test --release test_performance

Troubleshooting

Command not found

# Ensure binary is in PATH
echo $PATH
which ph

# If using cargo install
cargo install --list | grep prompthive

Clipboard not working

# If terminal, it copies automatically
ph use api  # Copies to clipboard

# Or pipe explicitly
ph use api | pbcopy  # macOS
ph use api | xclip  # Linux

Shell completions not working

# Regenerate completions
ph completion bash > ~/.bash_completion.d/prompthive
source ~/.bashrc

🤝 The Complete Terminal AI Stack

PromptHive + CalmHive = Ultimate AI Development Workflow

PromptHive handles instant prompt access (8ms), while CalmHive handles background processing, voice control, and smart execution management.

# Install the complete stack
cargo install prompthive
npm install -g @calmhive/calmhive-cli

# Use together for powerful workflows
calmhive afk "$(ph use essentials/refactor)" --iterations 20  # Background processing
calmhive voice "$(ph use essentials/debug)"                  # Voice control
ph use essentials/commit | calmhive process --smart-retry    # Smart execution

Perfect Division of Labor:

PromptHive: Instant prompt management, universal AI tool compatibility
CalmHive: Background processing, usage limit handling, voice control

Learn more: prompthive.sh | calmhive.com

Performance

PromptHive is engineered for sub-15ms performance across all operations. Built with Rust for maximum efficiency and reliability, it handles thousands of prompts without slowing down your workflow.

Tested on version 0.2.8: All core operations (new, use, show, edit, ls, find) complete in under 15ms on modern hardware.

Join the Community 🐝

🌟 100% Open Source & Free

✅ Unlimited local prompts
✅ Lightning-fast performance (<15ms)
✅ Compose & chain prompts
✅ Cross-platform support
✅ Text cleaning & formatting
✅ Magic link authentication
✅ Cloud sync across devices
✅ Private prompt banks
✅ Team collaboration & sharing
✅ Prompt sharing features
✅ Advanced analytics & insights
✅ Community registry access
✅ Self-hostable infrastructure

"Professional prompt management for everyone"

🤝 Contributing

Report bugs and request features
Submit prompts to the community registry
Contribute code improvements
Help with documentation
Join discussions and share workflows

Visit our GitHub repository to get involved!

📚 Resources & Links

PromptHive.sh - Official website with documentation and installation guides
CalmHive.com - Perfect companion for background AI processing and voice control
GitHub Repository - Source code, issues, and contributions
Crates.io Package - Rust package with version history
Blog - Terminal velocity philosophy and AI development insights

"Prompts that just work."

prompthive 0.2.8

PromptHive 🐝

TL;DR

🚀 Installation

Cargo (Rust) - Available Now ✅

Quick Install Script

Shell Completions

Bash

Zsh

Fish

Quickstart

"But Why Not Just..."

...use text files in ~/prompts/?

...use shell aliases?

...use Neovim with snippets?

...use ChatGPT/Claude history search?

...use GitHub Gists or company wikis?

The Real Differentiator: Workflow Velocity

Commands (v0.2.8)

🚀 Power User Examples

Built-in Template Variables

Version Control - Never Lose a Good Prompt

Diff & Merge - Collaborate on Prompts

Web Dashboard - Visual Analytics

🎯 Unified I/O Design - Smart Defaults

Text Transformation Commands (use, show, clean)

Query Commands (ls, find)

Universal I/O Flags

Smart File Operations

Composable Flags

Design Philosophy

Configuration

Disable Logging

Other Environment Variables

Creative Usage Patterns

🎯 Daily Journaling & Reflection

🧠 Learning & Study Assistant

💼 Meeting & Communication Templates

🎨 Creative Writing & Content

🔧 DevOps & Operations

📊 Data Analysis & Research

🎮 Personal Productivity

🚀 API & Integration Patterns

Performance

Partial Matching (Like JJ)

The Magic: Composition

Real-World Usage

Daily Developer Workflows

Advanced Workflows

🧙 Workflow Automation Examples

🔰 Shell Pipes Primer (for Terminal Newbies)

The Basics

With PromptHive

When to Use What

🏦 Prompt Banks - Instant Productivity

Built-in Prompts Ready to Use

Create Your Own Banks

Sharing Banks with Your Team

Real Workflows with Banks

Building PromptHive with Itself

File Structure

Prompt Format

Why PromptHive?

Features

Core Features (Production Ready)

Advanced Features (All Included - Free & Open Source)

Development Build

Troubleshooting

Command not found

Clipboard not working

Shell completions not working

🤝 The Complete Terminal AI Stack

Performance

Join the Community 🐝

🌟 100% Open Source & Free

🤝 Contributing

📚 Resources & Links