SAM | Turn Web Projects Into Automated Social Content

April 14, 2025
Nolan, FunkPd. SAM; Social Automation Machine; kills the Instagram grind. It’s a service at sam.funkpd.com, not a plugin. Sign up, sync your account, and it posts; quotes, photos; while you work. I built it to stop wasting time on captions. Runs FunkPd’s feed, helps welders, bakers stay visible. Researchers: it’s a lean AI pipeline for content and publishing. If you need a broader CRM and automation service, this is the same playbook.

The automation philosophy behind SAM matches the playbook in our AI-as-better-spreadsheet post ; tight loops, clear constraints, no fantasy.

Layman Explanation

Problem

Social media’s a slog. Businesses need it. You’re too busy welding, baking, building. No time for photos or captions. Skip it, your shop’s invisible.

Agitate

Blank Instagram screen. No ideas. An hour gone for one post. “Experts” sell $1,000 plans. You need customers, not likes. Rivals post daily, stealing leads. You’re stuck, hating every second.

Solution

SAM’s a service. Hit sam.funkpd.com, register, sync Instagram. AI makes posts, schedules them. Live in ~30 minutes. Your feed runs; shows your work; while you swing a hammer. Pro tier: up to 25 posts/day, forever. X, LinkedIn, Pinterest soon. No apps, no tech hassle. Saves time. Gets eyes on you.

For Researchers

SAM; Social Automation Machine; is a hosted Instagram automation service at sam.funkpd.com (tba), strictly Instagram-focused as of April 2025, with integrations for X, LinkedIn, or Pinterest planned for Q3 2025. Users; tradespeople, small businesses; register via reCAPTCHA-protected forms, OAuth-sync Instagram (storing Access Tokens/Business IDs securely), and trigger posts through a PHP/WordPress dashboard (SAM-dash). The Python-based SAM-core fetches verified data (e.g., ZenQuotes API), tags it with user-specific metadata using LLMs (Groq for speed, Ollama local fallback), and generates posts via HTML/CSS templates (imgz.py, Playwright, ~6s/image) or AI generated images. Structured prompts cut LLM hallucinations, hitting 99-100% output accuracy. Modular pipelines (users//modules/) isolate content logic, stored in SQLite (per-user-per-module DBs: raw, processed, pending, published). Async tasks (fetching, scoring, rendering) ensure ~10min/post delivery. SAM-api (PHP, JWT-secured) proxies requests; SAM-cast (internal WordPress) handles IG Graph API publishing (99% success post-retries). Lean stack: Python, PHP, SQLite, MySQL, minimal deps. Secure: HMAC core-to-cast, token auth, sanitized uploads. Pro tier will support 25 posts/day, multiple modules, team accounts. Legit, but AI inference, image rendering and DB scaling need work.

User Flow

1. See SAM’s Instagram post (@samfunkpd), ad, or word-of-mouth.
2. Register at sam.funkpd.com (bot protection included).
3. Sync Instagram via secure auth; SAM handles credentials.
4. Generate a post; it’s live in ~30 minutes.
5. Pro tier: schedule 25 posts/day, more platforms coming.

Core Components

• SAM-core (Python) : Content engine. AI generates text/images, manages users/modules. Async tasks, lightweight databases per module. Chains fetching, scoring, building, sending.
• SAM-api (PHP) : Secure gateway. Authenticates requests, proxies to core. Handles user and content actions.
• SAM-cast (Internal) : Our bridge. Takes core output, stores posts, publishes to Instagram via API. Custom storage, scheduled jobs.
• SAM-dash (Frontend) : User portal. Clean UI for signup, sync, post triggers. No user tech needed.

Key Features

• AI-Driven Precision : Fetches verified data, tags it with customer context, and crafts posts via structured LLM prompts. Cuts hallucination risks, ensures output accuracy.
• Modular Flexibility : Per-user content pipelines (e.g., quotes, photos) swap seamlessly, no core rewrites.
• Async Efficiency : Non-blocking fetches, renders, and posts for high throughput.
• Robust Security : Signed requests, token auth, bot defenses keep data tight.
• Lightweight Design : Minimal dependencies, small footprint, built to scale.

Tech Stack

// Backend: Python (image rendering, AI calls)// API: PHP// Frontend: Web UI, JavaScript// Storage: Lightweight DBs// APIs: Instagram, AI services// Security: Tokens, signatures

Workflow (Sample)

# Core pipelineasync def run():    raw = await fetch_data()    scored = await rank_data(raw)    posts = await build_posts(scored)    await send_posts(posts)

1. Fetch raw content.
2. AI ranks it.
3. Build posts (text + images).
4. Send to internal system.
5. Schedule and post.

Challenges

• AI output: 30% errors early, now ~10%.
• Image rendering: 6s each, down from 10s.
• API reliability: 99% post success after retries.
• Logs: Caught task conflicts early.

Results

Three months:

• FunkPd: 0 manual posts, daily feed.
• Engagement: +1000% (2 to 137 likes).
• Saved: 22 hours ($2,200).
• Post time: 47s vs. 45min.

Tradespeople: Your work’s seen. Researchers: Study the pipeline; AI, async, modular.

Next

• Images: 6s to sub-3s? Tips?
• Captions: Less bot, more grit. Prompt ideas?
• Scale: Queue options for 1,000+ users?
• Testers: 10 businesses for sam.funkpd.com. DM me.

SAM works. Saves an hour, it’s worth it.
; Nolan, FunkPd

Try it: sam.funkpd.com. Questions? FunkPd.com or X @funk_pd.