Launching Q3 2026 — Join the waitlist

Stop burning tokens on
tasks a script can handle.

Proactive Runner decomposes your goals into a journey map, classifies each step as deterministic or adaptive, and routes them to the right execution engine.

Scripts for what's predictable. AI for what isn't. 60%+ cost savings.

~/project
$ proactive plan "Find CTOs of 50 newest YC companies and get their emails"
⟳ Decomposing goal into journey map...
✓ 7 steps identified
🔧 4 DET🧠 3 ADPT~$1.56~12 min
$ proactive run
✓ Step 1: Get YC list DET 0.3s
✓ Step 2: Find websites DET 1.2s
⟳ Step 3: Locate team pages ADPT 23/50 █
60%+
Cost Savings
CLI-First
Developer Experience
Open Source
Free Forever (Local)
Scroll

How It Works

From natural language to executed results in three steps. No workflow builder. No drag-and-drop. Just describe and run.

💬
STEP 01

Describe Your Goal

Tell Proactive what you want in plain English. No config files, no YAML, no learning a new DSL.

$ "Find CTOs of 50 YC S25 companies and get their emails"
🧠
STEP 02

AI Maps & Classifies

The AI decomposes your goal into a journey map and classifies each step as Deterministic (script/API) or Adaptive (AI agent).

🔧 Get YC List
→
🔧 Find Websites
→
🧠 Team Pages
→
🧠 Identify CTOs
→
🧠 Find Emails
→
🔧 Verify
→
🔧 Export
🚀
STEP 03

Execute Anywhere

Run locally in Docker (free) or on cloud runners (paid). Same Dockerfile, same results. Watch progress in real-time.

$ proactive run --cloud

The Automation Paradox

Today's tools force a false choice. You either script everything manually or throw tokens at every step. Neither is smart.

⚙️

Traditional Automation

Zapier, Make, n8n

  • ✕Rigid — breaks when websites change
  • ✕Manual setup for every step
  • ✕Zero intelligence, can't handle exceptions
  • ✕Works until it doesn't
Low cost, low capability
🤖

Full AI Agent

Raw Claude Code, Devin

  • ✕Expensive — $0.05–$2.00+ per task
  • ✕Unpredictable execution time
  • ✕Overkill for deterministic operations
  • ✕Every step burns tokens, no optimization
High cost, high capability (but wasteful)
↓
⚡

Proactive Runner

Intelligent routing — the right tool for each step. Scripts handle what's predictable (fast, cheap, reliable). AI handles what needs judgment (flexible, smart). You pay only for intelligence where it actually matters.

🎯 Smart routing💰 60%+ cheaper📈 Self-improving

Real Use Cases

Not hypotheticals. These are actual workflows with real cost breakdowns.

📧

YC CTO Email Finder

Find CTOs of 50 YC S25 companies, get verified emails. Fan-out across companies, parallel execution.

🔧 4 DET🧠 3 ADPT~12 min
Full AI: $4.50
Proactive: $1.56
Save 65%
📊

Competitor Pricing Monitor

Monitor 5 competitors' pricing pages daily. Smart change detection triggers AI classification only when needed.

🔧 5 DET🧠 2 ADPT~1.5 min/day
Full AI: $15/mo
Proactive: $3/mo
Save 80%
📱

Cross-Platform Publishing

Take blog posts, create platform-optimized variants for Twitter/X, LinkedIn, and Dev.to. Publish automatically.

🔧 5 DET🧠 2 ADPT~2 min
Full AI: $2.00
Proactive: $0.25
Save 87%
📄

Contract Clause Extraction

Process 30 vendor contracts — extract payment terms, termination clauses, liability caps. Flag anomalies.

🔧 4 DET🧠 2 ADPT~14 min
Full AI: $12.00
Proactive: $2.90
Save 76%

CLI Preview

Everything starts from the terminal. Plan, run, monitor — all from a single command.

proactive — ~/yc-outreach
# Initialize a new project from a template
$ proactive init --template yc-outreach
✓ Project initialized from 'yc-outreach' template
# Describe your goal — AI generates a journey map
$ proactive plan "Find CTOs of 50 YC S25 companies, get verified emails"
⟳ Decomposing goal...
✓ Journey map: 7 steps (4 DET, 3 ADPT)
Est. cost: $1.56 | Est. time: ~12 min
Run "proactive plan --edit" to modify
# Execute locally in Docker (free)
$ proactive run
▶ Running in local Docker container...
Step 1/7: Get YC list ·········· ✓ DET 0.3s
Step 2/7: Find websites ········ ✓ DET 1.2s
Step 3/7: Team pages ··········· ✓ ADPT 4m12s
Step 4/7: Identify CTOs ········ ✓ ADPT 3m05s
Step 5/7: Find emails ·········· ✓ ADPT 5m22s
Step 6/7: Verify emails ········ ✓ DET 0.8s
Step 7/7: Export CSV ··········· ✓ DET 0.1s
✓ Complete — 48/50 CTOs found | $1.43 spent | 12m54s
# Or execute on cloud runners (paid, no Docker needed)
$ proactive run --cloud
# Check status of running executions
$ proactive status
proactive initCreate a project from a template
proactive planGenerate journey map from a goal
proactive runExecute locally in Docker
proactive run --cloudExecute on cloud runners
proactive statusCheck progress & cost
proactive templatesBrowse community templates

Cloud Runners

Run locally for free. Scale to cloud when ready.

Same Dockerfile, same results. Cloud runners mean no Docker installed locally, burst compute for parallel fan-out, and scheduled execution.

🌱

Starter

$0.01/min
CPU0.5 core
RAM512MB
Simple scripts, API calls
Popular
⚡

Standard

$0.03/min
CPU2 cores
RAM4GB
Browser automation, data processing
🚀

Performance

$0.05/min
CPU4 cores
RAM8GB
Heavy compute, batch processing
🎮

GPU

$0.15/min
CPU2 cores + GPU
RAM16GB
ML inference, image generation

Local mode is always free. Just have Docker installed. Your data stays on your machine.

🐙

Open Source at the Core

The CLI is open source. Always will be. Run locally forever at zero cost.

Cloud runners pay the bills. The core engine — journey mapping, classification, local Docker execution — is yours to inspect, fork, and contribute to.

Star on GitHub
MIT License
Node.jsTypeScriptDockerCommander.jsClaude CodeOpenClaw
⚡

Be first to run proactively.

Join the waitlist for early access. CLI beta launching Q3 2026.

No spam. Just launch updates.

🔒 No credit card required
🐳 Local mode always free