ShareCaster Studio
Florida ShareCaster Labs LLC • sharecaster.com
FLORIDA SHARECASTER LABS LLC • SOVEREIGN MULTI-AGENT CONTROL PLANE

Pluggable Agent Intelligence. Cryptographic Human Command.

Built from first principles on a high-concurrency Rust/Axum control plane and Python/FastAPI orchestration layer. Run our built-in reflection loop, hook up your own HMAC-SHA256 Webhook Agents or MCP 2026-07-28 Tool Servers, and route tasks across Google Gemini, Anthropic, OpenAI, and local Ollama—locked behind a hard database approval gate.

Launch Live Sandbox
Active Agent Mode
BUILT-IN REFLECTION
Tier-1 Ingestion
OLLAMA • $0.00
Tier-3 Critique
GEMINI-4-CARBON
Approval Gate
AWAITING_APPROVAL
EXTENSIBLE AGENT OPERATING SYSTEM • HMAC-SHA256 & MCP 2026-07-28

Custom External Agents & MCP Server Bridge

Treat intelligence workers as pluggable microservices. Switch seamlessly between our built-in Python/FastAPI reflection pipeline, your own HMAC-SHA256 signed webhook agents (LangChain, CrewAI, n8n, custom Python/Rust), or a Model Context Protocol (MCP 2026-07-28) tool server—while every callback still passes through our hard approval gate.

Zero-Trust Callback Policy
HARD GATE ENFORCED
AGENT REGISTRY & HANDSHAKE
● VERIFIED (24 ms)
Route Callback Through Hard Approval Gate

External agent outputs pause at AWAITING_APPROVAL before social API dispatch.

Header: X-ShareCaster-Signature-256: sha256=8f4e...c91a Spec: HMAC-SHA256 / Rust Axum

              
Supported Frameworks: LangChain • CrewAI • n8n • MCP SDK • Custom Rust/Python OAuth 2.1 + Client ID Metadata Ready
BYOK & DYNAMIC MODEL POWER ROUTER WORKBENCH

Route Every Task to Its Optimal Intelligence Tier

Plug in your own Google Gemini, Anthropic, OpenAI, OpenRouter, or local Ollama endpoints. FastAPI automatically matches task complexity to model power—keeping 80%+ of token volume at zero or near-zero cost.

Blended Savings vs All-Frontier
86.4% ($342/mo saved)
TIER 1 • INGESTION & SCORING $0.00 / 1M

High-Volume Trend Scoring

Agent: trend-scout

Deduplicates webhooks, normalizes feeds, and runs velocity scoring before any drafting tokens are spent.

TIER 2 • 1M-TOKEN DRAFTING $1.50 / 1M

Standard Drafting & Repurposing

Agent: content-synthesizer

Ingests full multi-hour transcripts in a 1M-token pass and fragments them into platform-native threads, posts, and scripts.

TIER 3 • CRITIC & MEDIA $2.00 / 1M

Frontier Critic, Refiner & Media

Agents: critic-refiner + media-synthesizer

Audits brand alignment, sharpens viral hooks, and synthesizes thumbnail/visual prompts before locking at the approval gate.

AES-256 ENCRYPTED KEY VAULT & 1-TOKEN LATENCY PROBE

Verify credentials and measure live round-trip latency across Google Gemini, Anthropic, OpenAI, OpenRouter, and your local Ollama LAN node.

GOOGLE_GEMINI
● VERIFIED 114 ms
ANTHROPIC
● VERIFIED 182 ms
OPENAI
● VERIFIED 146 ms
OPENROUTER
● VERIFIED 161 ms
OLLAMA_LAN LOCAL
● GPU READY 8 ms
LIVE TASK ESCALATION SIMULATOR

Compare a routine feed scrape executed 100% on Tier 1 versus escalating a viral campaign across Tier 2 and Tier 3.

1. TIER-1 INGESTION $0.0000
Local Ollama (Llama 3 / Mistral)
✓ Filtered 210 signals & scored velocity
2. TIER-2 SYNTHESIS SKIPPED
Google Gemini (gemini-4-hydrogen 1M / 3.8-flash)
Bypassed — below viral velocity threshold
3. TIER-3 CRITIC & MEDIA SKIPPED
Google Gemini (gemini-4-carbon + nano-banana-2.1)
Zero frontier tokens burned on routine poll
Routing Decision: Routine feed poll executed 100% on Tier 1. Run Spend: $0.0000 (Saved $0.048 vs Frontier)
BUDGET CIRCUIT BREAKER, PGVECTOR RAG & STRUCTURED OUTPUT GUARDRAILS

Zero Runaway API Bills. Zero Brand Drift.

BUDGET CIRCUIT BREAKER ● ARMED & HEALTHY

Daily Workspace Spend Cap ($1 – $100/day)

Tracks real-time token burn in PostgreSQL via SQLx. If a viral spike hits your daily cap, the router automatically downgrades Tier 3 → Tier 1 (Ollama / gemini-3.5-flash-lite).

Daily Spend Cap: $15.00 / day
$1/day $50/day $100/day
Current Ledger Burn: $2.40 16% of cap
Fallback Policy: Tier 3 → Tier 1 Auto-Downgrade

pgvector Brand Voice RAG

gemini-embedding-2

Indexes past winning posts, style rules, and banned-phrase lists inside PostgreSQL pgvector, injecting top-3 cosine matches into every draft and critique cycle.

Vector Index: 1,420 winners (768-dim) ● ACTIVE (Top-3 Inject)

Strict PublishPayload JSON Schema

● ENFORCED

Validates both built-in LLM outputs and external webhook/MCP agent callbacks against strict Rust PublishPayload JSON schemas with automatic retry before database persistence.

Validation Engine: response_schema + Pydantic v2 0 Malformed Payloads
Integrated Agent & BYOK Telemetry Sandbox

Live SSE Reflection & Hard Approval Gate

Stream live execution telemetry reflecting your active Agent Mode (Built-In, HMAC Webhook, or MCP 2026-07-28) and BYOK Power Tiers, then execute cryptographic sign-off.

Streak: 24 Days
npub1sc9...4m2x
Presets:
sse://api.sharecaster.com/v1/pipeline/stream
bundle: #sc-9104
SQLx State: AWAITING_APPROVAL
Operator Sig: 0x8c2f...4e9a (Pending Sign-Off)
Chars: 246 / 280 • Hook Verified Schema Valid

                
Approve & Drip Cadence
Retention, Cryptographic Provenance & Adaptive Learning

Closed-Loop Sovereign Intelligence

SQLx Streaks

SQLx Publishing Streaks

PostgreSQL SQLx tracks continuous operator sign-offs and trend-capture cadence, unlocking priority Tokio worker scheduling and streak milestones.

Current Streak 24 Days • 1.5x Priority
NIP-07 + IPFS

Nostr NIP-07 / IPFS Audit Receipts

Sign in with Nostr NIP-07 public keys and pin immutable provenance receipts to IPFS / Arweave on every operator sign-off—proving human review over every bundle.

Audit Receipt Immutable CID Verified
performance-tracker

Adaptive Prompt Weighting Loop

Post-publish Rust pollers ingest live CTR and engagement telemetry back into PostgreSQL, re-ranking pgvector exemplars and updating system prompt weights automatically.

Hook Adaptation +38.4% CTR Lift
Unlimited BYOK & Custom Agents Across All Tiers

Self-Host Forever or Deploy Hybrid Cloud

SELF-HOSTED CORE

Sovereign Self-Hosted

Complete Docker Compose stack on your own server or workstation with zero recurring fees.

$0
/ month Forever free open-core
  • Unlimited BYOK keys (Gemini, Anthropic, OpenAI, OpenRouter, Ollama)
  • Full HMAC Webhook & MCP 2026-07-28 custom agent bridge
  • PostgreSQL + pgvector RAG & Budget Circuit Breaker
  • Hard AWAITING_APPROVAL gate & Nostr NIP-07 auth
Most Popular • Hybrid Cloud
CREATOR & OPERATOR

Creator Pro

Managed Rust/Axum control plane with hybrid local Ollama tunnel, mobile approval bots, and staggered drip scheduling.

$49
/ month Billed monthly
  • Unlimited BYOK keys + Custom HMAC & MCP Agents
  • One-tap Telegram & Slack mobile quick-approve webhooks
  • Approve & Drip 7-day staggered campaign scheduler
  • Closed-loop Post-Publish Analytics & IPFS audit receipts
MULTI-TENANT TEAMS

Agency / Syndicate

Isolated client workspaces, per-client AES-256 BYOK vaults, and dedicated external agent fleets.

$149
/ month Billed monthly
  • Unlimited BYOK keys & isolated per-client AES-256 vaults
  • Multi-endpoint MCP 2026-07-28 & HMAC agent registries
  • Per-workspace Budget Circuit Breakers & Stripe billing
  • Nostr NIP-07 role-based operator sign-off controls
Configuration updated.