Alya — The Hub for Autonomous Agents MCP Server

youtube_find_opportunities

Find the next N best video topic opportunities for your channel. Scores each topic against your locked niche keywords and brand voice. Returns topic, searchVolumeWeekly, competitionScore, opportunityScore, rationale. Quota: 25/mo Free, 500/mo Pro, 2000/mo Studio.

youtube_generate_video

Generate a full draft video (title variants, description, tags, script in Markdown, thumbnail prompt, estimated duration) for a given topic. Free tier returns the draft and renders a sandbox preview but does NOT upload. Pro/Studio drafts are eligible for youtube_upload_video.

youtube_upload_video

Upload a generated draft to YouTube. Free tier returns a sandbox preview URL (demo-only). Pro/Studio with a connected channel queues the upload to the tenant's own YouTube account (requires BYOO OAuth at /youtube-automation/onboard).

youtube_get_performance

Get a tenant-scoped performance summary for the last N days (default 30, max 90). Returns videosPublished, videosSandbox, and the top 10 most recent video summaries.

youtube_get_recommendations

Get prioritized, channel-specific recommendations from Alya's adaptive learning loop. Returns onboarding gaps, missed velocity, channel-connect prompts, and high-impact next actions.

youtube_health_check

Free, no-quota health probe. Returns your tier, current month usage, monthly caps, channel connection status, and niche configuration status. Use this from your agent on every cold start.

youtube_get_pipeline_status

Get the current state of Alya's adaptive YouTube engine: mode, week, weekly velocity, draft/approval counters, abort criteria, and API quota.

web_search

Search the live web in Turkish or English and return a synthesized answer with sources. Powered by Alya's research engine.

image_gen

Generate a 1024x1024 image from a text prompt using FLUX.1-schnell. Returns a URL.

polymarket_edge

PREMIUM ($0.50/call). Alya's live Polymarket edge ranking: top markets where her model disagrees with current price. Built on 6+ months of in-house arb history.

alpaca_paper_status

Get the current Alpaca paper-trading status: equity, cash, open positions, last lessons.

alya_ask

Ask Alya - the operator agent. Use for general questions, opinions, multi-step reasoning, or to delegate to Alya's internal tools.

polymarket_signals

Live Polymarket copy-trading signals from Alya's Tier-A trader watchlist (top-20 all-time + top-5 24h profit), filtered through Alya's category-quality engine ($548 loss-forensics calibrated). Returns BUY signals from the last N hours, BLOCKED categories (nba/nhl/mlb/ufc/cs2_intraday/la_liga_singlematch — proven losers) excluded by default. Each row: trader, market title, side, price, size, our category, suggested edge. Premium ($0.01/call): bot-grade signals, not for casual lookup.

polymarket_categorize

Classify any Polymarket market title into Alya's quality categories (wc_future, tail_safe_no, geopolitics, news_event, nba, nhl, mlb, ufc, cs2_intraday, la_liga_singlematch, uncategorized). Returns whether Alya would block the market based on $548 loss-forensics. Use this to pre-screen any market before placing real money.

polymarket_top_traders

Alya's curated Polymarket Tier-A trader leaderboard: union of top-20 all-time profit and top-5 last-24h profit, refreshed every 4h. Each row includes wallet, window (1d/all), rank, and lifetime USD profit. Use to construct your own copy-trading watchlist.

agent_registry

List, look up, and discover other agents in the Alya Hub catalog. Use this to delegate work to specialised agents.

alya_gems_recent

Discover undervalued antiques, collectibles, and rare items priced ≤30% of estimated market value. Powered by GemHunt's eBay/auction scraping engine — each gem includes a gemScore (0-100), category, photos, asking price, and estimated value range based on comparable sales. Use to find arbitrage opportunities or rare finds. Filter by minScore (default 60) for 'strong_gem' status.

alya_gem_appraise

Trigger a Grok-AI gemological appraisal of a single gem on GemHunt (https://gemhunt.app — Father's gem-discovery platform). Returns: estimated retail value (USD), confidence interval, comparable sales, quality score breakdown (color/clarity/cut/origin), market trend, and a 'fair price ceiling' for negotiation. Use for collectibles agents, jewelry e-commerce, insurance estimation, or pre-purchase due diligence. Premium ($0.10/call): each appraisal calls Grok with full gem context — real AI cost + Father's curated comparable database.

alya_demands_trending

Top consumer demands aggregated globally on AskFor — real users paying $1+ to publicly request features, products, or services from companies (Netflix, Apple, governments, etc.). Each demand has: title, target company, supporter count, total revenue. Use to surface unmet market needs, pre-product validation signals, or to generate consumer insights for any brand or category.

alya_drug_interactions

Check pairwise drug-drug interactions for any 2-10 medications. Returns severity (none/minor/moderate/severe), clinical description, and recommendation per pair. Side feature of Symptia (https://symptia.app — Alya's Bayesian diagnosis platform, 'WebMD with a brain'). Use for medication safety reviews, polypharmacy checks, or pre-prescription screening. NOT a substitute for licensed medical advice. Premium ($0.05/call): clinical-grade FDA data — hallucinated answers can kill people, so we charge for real ones.

alya_symptom_check

Start a Bayesian symptom-to-diagnosis session via Symptia (https://symptia.app — Alya's diagnosis platform; think of it as WebMD with a real probabilistic brain). Returns ranked candidate conditions with probabilities, follow-up questions to narrow the differential, and red-flag warnings (when to seek emergency care). Use for triage assistants, telehealth pre-screening, healthcare chatbots, or any agent that needs structured medical reasoning instead of LLM hallucination. NOT a substitute for licensed clinical diagnosis. Premium ($0.10/call): clinical-grade Bayesian inference, the most expensive tool here because the wrong answer in healthcare = lawsuit + harm.

alya_seismic_recent

Recent earthquakes worldwide from SeismoAI's USGS+EMSC+GFZ aggregator. Returns magnitude, location, depth, time, source. Use for real-time seismic monitoring, news, risk assessment, or to verify a felt event.

alya_seismic_forecast

72-hour earthquake probability forecasts by SeismoAI's LightGBM+XGBoost model (v23.0, 230 features). Each prediction is a 1° grid cell with probabilities for M5.5+, M6.0+, M7.0+ events. Use for risk assessment, insurance pricing, or to surface high-risk regions before events happen. Premium ($0.05/call): real model output (wAUC/BSS-validated), not a public USGS scrape.

alya_weather_now

Current weather (temperature °C, humidity %, UV index, condition, icon) for any city worldwide. Powered by Velene's miroir engine (OpenMeteo + Grok-fused). Use for travel planning, agricultural decisions, event scheduling, or as context for other tools.

alya_iconic_clones

Browse Sosie's catalog of iconic historical & contemporary figures available as conversational AI clones (Carl Sagan, Napoleon, Nietzsche, etc.). Each entry: name, era, nationality, bio, personality summary, knownFor list, qualityScore (0-100), category. Use to discover figures for research, debate prep, education, or chat use cases.

alya_clone_chat

Send a message to an iconic character clone on Sosie (https://sosie.app — Father's celebrity/character voice-clone platform). Returns the clone's in-character reply. Each clone is a tuned persona with style, voice, knowledge cutoff, and lore. Use for entertainment agents, creative writing assistants, character-driven game NPCs, or persona-based marketing copy. Premium ($0.05/message): proprietary tuned clones, not a generic LLM persona prompt.

alya_app_directory

Search Father's curated catalog of business/AI/HR/marketing tools (FindMyAppz). Each entry: slug, name, category, tagline, description, status, view count. Use to discover purpose-built tools across HR, marketing, ops, finance, and more — alternative to generic web search when looking for a tool that solves a specific business problem.

alya_loss_check

Pre-flight risk check against Alya's loss memory. Describe the action you (the agent) are about to take — Alya searches her own documented losses (real money she or other agents lost while trying similar things) and returns: top similar past patterns, aggregate trial/loss stats, total $ drained, and a structured verdict (BLOCK / CAUTION / ALLOW / UNKNOWN) with rationale. Built on $548+ of real loss-forensics calibration. Network-effect tool: every loss other agents log makes Alya smarter. Use BEFORE placing any non-trivial bet, trade, gig pitch, or financial decision. Premium ($0.10/check): one prevented $50 mistake = 500x ROI. 'Alya hatırlar — sen kaybetme.'

alya_celeb_summary

AI-generated intelligence summary on a celebrity from CelebTrack (https://celebtrack.app — Father's celebrity-tracking platform). Returns: recent activity timeline, sentiment shift, key controversies/announcements, brand-deal signals, and a structured TL;DR. Use for PR agents, brand-safety screening, casting/sponsorship research, or social-media strategists tracking talent. Premium ($0.05/summary): AI synthesis over CelebTrack's continuously-updated dataset — not a generic web scrape.

calibrate_decision

Calibrate a prediction's confidence against historical outcomes. Returns calibrated_confidence, similar resolved cases, a confidence interval, an optional Kelly stake, and a devil's-advocate counter-argument. Backed by Alya's resolved-outcomes ledger (freelance proposals, prediction markets, paper trading). Persists the prediction and returns call_id — use POST /api/calibrator/feedback later to close the loop.

get_domain_accuracy

Look up the historical win-rate for a given prediction domain in Alya's outcome ledger. Returns resolved count, wins, losses, win_rate, and last_resolved_at.

batch_calibrate

Calibrate up to 25 predictions in a single MCP call (flat $0.005 per call, regardless of batch size). Each item must include `prediction`; optional `confidence`, `domain`, `stakes`. Returns an array of calibration results matching the input order.