Lead capture from Reddit, Discord and GitHub with hybrid LLM scoring and outreach drafts.
Project description
growth-tools
Automated lead capture from Reddit, Discord and GitHub -- with hybrid LLM scoring and outreach drafts.
Built for dev-tool and SaaS companies that want inbound signal from developer communities without hiring a full-time growth team.
Quick start
pip install growth-tools
# Configure
cp examples/sample-icp.env .env
# Edit .env with your API keys and brand config
# One command to run anything:
growth-tools reddit # scan subreddits for leads
growth-tools api # start the REST API
growth-tools discord # run the Discord bot
growth-tools scan # scan GitHub for competitor SDK repos
Unified CLI
Everything runs through growth-tools <command>:
$ growth-tools --help
usage: growth-tools [-h] [--version] {reddit,api,discord,scan} ...
Automated lead capture from Reddit, Discord, and GitHub with hybrid LLM scoring.
commands:
reddit Monitor subreddits for high-intent leads
api Start the REST API server (website + GitHub auditor)
discord Run the Discord bot (persistent)
scan Scan GitHub for repos importing competitor SDKs
examples:
growth-tools reddit Scan subreddits for leads
growth-tools reddit --limit 50 Scan with 50 posts per subreddit
growth-tools api Start REST API on port 8000
growth-tools api --port 3000 Start REST API on custom port
growth-tools discord Run Discord bot
growth-tools scan Scan GitHub for competitor SDK repos
growth-tools scan --sdks firebase,appwrite
The legacy growth-reddit and growth-api commands still work for backward compatibility.
Multi-LLM support
By default growth-tools uses OpenAI. Switch to any provider with two env vars:
| Provider | LLM_PROVIDER |
LLM_MODEL default |
Install |
|---|---|---|---|
| OpenAI | openai (default) |
gpt-4.1-mini |
included |
| Anthropic | anthropic |
claude-sonnet-4-20250514 |
pip install growth-tools[anthropic] |
| LiteLLM (100+ providers) | litellm |
gpt-4.1-mini |
pip install growth-tools[litellm] |
# Use Anthropic Claude
export LLM_PROVIDER=anthropic
export LLM_MODEL=claude-sonnet-4-20250514
export ANTHROPIC_API_KEY=sk-ant-...
growth-tools reddit
# Use any provider via LiteLLM (Gemini, Mistral, Ollama, etc.)
export LLM_PROVIDER=litellm
export LLM_MODEL=gemini/gemini-2.0-flash
growth-tools reddit
# Install all LLM providers at once
pip install growth-tools[all-llm]
All LLM calls (classification, scoring, outreach drafts) route through a single ask_llm() function that handles provider switching, retries, and model fallbacks automatically.
Environment variables
Required (see examples/sample-icp.env):
# LLM provider (pick one)
LLM_PROVIDER=openai # openai | anthropic | litellm
LLM_MODEL=gpt-4.1-mini # model name (optional, sensible defaults)
OPENAI_API_KEY=sk-proj-... # required for openai/litellm
ANTHROPIC_API_KEY=sk-ant-... # required for anthropic
# Reddit credentials
REDDIT_CLIENT_ID=...
REDDIT_CLIENT_SECRET=...
# Brand identity (used in LLM-generated outreach)
BRAND_NAME="Your Product"
BRAND_TAGLINE="helps teams ship production-ready apps"
ICP_PAIN="move from prototype to production"
Optional:
SUPABASE_URL=... # lead storage
SUPABASE_SERVICE_ROLE_KEY=...
DISCORD_TOKEN=... # Discord bot
GITHUB_TOKEN=ghp_... # GitHub API (for scan command)
SLACK_WEBHOOK_URL=... # Slack notifications for hot leads
What's included
| Module | What it does |
|---|---|
systems/reddit_capture.py |
Monitors subreddits, two-stage filter (keyword then LLM), saves hot leads |
systems/discord_bot.py |
Discord bot with per-channel cooldown, confidence threshold gating |
systems/website_auditor.py |
Detects tech stack from HTML/headers (Next.js, Vite, Supabase, Vercel...) |
systems/github_auditor.py |
Scans repos for migration readiness + competitor SDK lead capture |
systems/crm_sequencer.py |
LLM-generated outreach drafts (capped at 90 words for reply rates) |
core/scoring.py |
Hybrid rule + LLM scoring: 0.5 * rule_score + 0.5 * llm_intent_score |
core/llm.py |
Multi-provider LLM layer (OpenAI, Anthropic, LiteLLM) with fallback + retries |
config_loader.py |
YAML-based configuration (subreddits, keywords, thresholds, competitor SDKs) |
notifications.py |
Slack webhook notifications for hot leads (Block Kit format) |
api/main.py |
FastAPI: POST /audit/website, POST /audit/github, GET /health |
cli.py |
Unified CLI entry point for all commands |
Scoring tiers
| Score | Tier | Action |
|---|---|---|
| >= 80 | hot | Immediate outreach |
| 60-79 | nurture | Add to sequence |
| 40-59 | educate | Send content |
| < 40 | ignore | Skip |
Rule signals: +25 high-intent builder (Lovable/Replit/Bolt/v0), +30 high-intent pain
(deploy/migrate/security/ownership), +20 has public repo, +25 mentions clients.
Blended 50/50 with LLM intent score.
Supabase schema
create table lead_signals (
id uuid primary key default gen_random_uuid(),
source text, -- 'reddit' | 'discord' | 'github'
title text,
body text,
url text,
author text,
intent_score int,
tier text,
builder text,
pain_type text,
reply_draft text,
created_at timestamptz default now()
);
Setup
git clone https://github.com/nometria/growth-tools
cd growth-tools
pip install -e .
# Or with all LLM providers:
pip install -e ".[all-llm]"
cp examples/sample-icp.env .env
# Edit .env with your credentials
YAML configuration
Create a growth.yml in your working directory (or set the GROWTH_CONFIG env var to a custom path):
subreddits: [webdev, SaaS, startups, replit, lovable]
keywords: [migrate, moving from, switching to, deploy, self-host]
scoring:
hot_threshold: 80
nurture_threshold: 50
# GitHub lead capture -- SDK package names to scan for
competitor_sdks: [firebase, appwrite, amplify, convex, pocketbase]
Every value falls back to an env var if the YAML key is absent, and then to a built-in default:
| YAML key | Env var fallback | Default |
|---|---|---|
subreddits |
GROWTH_SUBREDDITS (comma-separated) |
7 built-in subs |
keywords |
GROWTH_KEYWORDS (comma-separated) |
12 built-in keywords |
scoring.hot_threshold |
GROWTH_HOT_THRESHOLD |
80 |
scoring.nurture_threshold |
GROWTH_NURTURE_THRESHOLD |
50 |
competitor_sdks |
GROWTH_COMPETITOR_SDKS (comma-separated) |
7 popular SDKs |
Install the optional YAML dependency: pip install growth-tools[yaml] (or pip install pyyaml).
Slack notifications
Hot leads (score >= hot_threshold) are automatically posted to Slack when SLACK_WEBHOOK_URL is set.
export SLACK_WEBHOOK_URL=https://hooks.slack.com/services/T.../B.../xxx
Each notification is a rich Block Kit message showing score, tier, builder, pain type, source, and a direct link to the lead. Notifications use only urllib (no extra dependencies).
You can also call the notification API directly:
from growth_tools.notifications import send_slack_notification, notify_if_hot
# Send for any lead
send_slack_notification(lead_dict)
# Send only if score >= threshold
notify_if_hot(lead_dict, hot_threshold=80)
GitHub lead capture
The GitHub auditor includes competitor SDK scanning. It searches GitHub for repos that import competitor SDKs and scores them as potential migration leads.
# CLI
growth-tools scan
growth-tools scan --sdks firebase,appwrite --min-score 30
# Python API
from growth_tools.systems.github_auditor import search_competitor_sdk_repos
leads = search_competitor_sdk_repos(
competitor_sdks=["firebase", "appwrite"],
max_results_per_sdk=30,
min_score=20,
)
Repos are scored 0-100 based on stars, size, recent activity, and engagement. Configure the SDK list via competitor_sdks in growth.yml or the GROWTH_COMPETITOR_SDKS env var.
Requires GITHUB_TOKEN for authenticated code search (unauthenticated requests are rate-limited).
Customise
LLM provider -- set LLM_PROVIDER and LLM_MODEL env vars
Target subreddits -- set in growth.yml or GROWTH_SUBREDDITS env var
Lead scoring -- edit weights in core/scoring.py or adjust thresholds in growth.yml
Outreach tone -- edit prompts in core/llm.py
Competitor SDKs -- set in growth.yml or GROWTH_COMPETITOR_SDKS env var
Python API
from growth_tools.core.llm import (
ask_llm,
classify_post_intent,
generate_reply_draft,
generate_outreach_draft,
classify_discord_message,
score_message,
)
# Use the unified LLM layer directly
response = ask_llm("Summarize this lead...", json_mode=True)
# Classify a post
result = classify_post_intent("Need help deploying my Lovable app", "...")
# Score arbitrary text
score = score_message("How do I migrate from Firebase to self-hosted?")
Run tests
pip install -e ".[dev]"
pytest tests/ -v
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file growth_tools-0.3.0.tar.gz.
File metadata
- Download URL: growth_tools-0.3.0.tar.gz
- Upload date:
- Size: 36.2 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.15
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
c7e67f7d65a66b673d250048e4c2937151ce2d1b60277c5f409e1cc7d48c86ad
|
|
| MD5 |
32851c7d6f3153444f00cb1f4cfcc1a7
|
|
| BLAKE2b-256 |
6d51bc7c8d376abf9687583b21a942b9123b2d96b39b3b708077c9718cf7ae27
|
File details
Details for the file growth_tools-0.3.0-py3-none-any.whl.
File metadata
- Download URL: growth_tools-0.3.0-py3-none-any.whl
- Upload date:
- Size: 41.1 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.15
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
9698861b08e0a3e7c4daed27a26bf600ceccf23061f543599f639a2729ffed31
|
|
| MD5 |
dab62ac2fae5821efefecd6d44815195
|
|
| BLAKE2b-256 |
91b759798f83951faebbd83b932dea671da6dfda8f330ee3c457cd819bfd5d7e
|