Discovery streams transform publicly available knowledge into structured intelligence for AI systems. Working within fair-use principles, we curate and synthesize content from press releases, podcasts, blog posts, SEC filings, research reports, and more — presenting it in machine-readable formats that AI can consume, reason about, and act upon.
Instead of inheriting the web's incentives built around SEO and ad performance, Discovery streams reorder what machines prioritize by surfacing verifiable, high-signal content. This gives AI systems a cleaner starting point for reasoning, free from the noise and blind spots of human-optimized search.
A story is a single piece of structured intelligence from a stream — one data point, one summary, one event. Think of it as one earnings call summary, one BLS employment report, one podcast episode breakdown, or one SEC filing update.
Stories are engineered for machines, not humans. They're designed to be consumed by AI systems via API or MCP at scale, with consistent structure, traceable sources, and machine-readable formatting that accelerates reasoning and decision-making.
Absolutely! Your credentials include a unique MCP url that you can connect to with your Pro ChatGPT accounts in Developer mode.
Yes! Claude has native MCP support. Connect your Synorb MCP server directly to Claude Desktop for seamless access to streams.
Recall windows determine how far back in time your AI systems can query for stories. This isn't storage — it's the depth of historical context available to your systems on-demand.
30 Day: Access stories from the last 30 days
60 Day: Access stories from the last 60 days
120 Day: Access stories from the last 120 days
Longer recall windows enable richer pattern recognition, trend analysis, and historical reasoning — essential for AI systems making informed decisions based on how situations have evolved over time.
Single Stream enables AI systems to query and analyze content within a single stream at a time. This is the standard way to consume Synorb data — perfect for most use cases where you need focused intelligence from one domain or topic.
Cross-Stream unlocks the ability to query and synthesize insights across multiple streams simultaneously — connecting dots between domains, identifying patterns across contexts, and building richer, more comprehensive intelligence. This is essential for complex analysis that spans topics, industries, or time periods.
Think of Single Stream as reading one book at a time, while Cross-Stream is like having a research library where you can connect ideas across multiple books in a single query.
Narrative streams turn oceans of numerical data into reasoning-ready text. Global data creation is at zettabyte scale and accelerating — humans cannot keep up, but machines can. Narrative streams translate complexity from datasets, sensors, and systems into precise language that AI can use for insight and decision-making.
Research streams deliver citation-ready analysis and reports assembled by reasoning agents. Each report preserves sources, states assumptions, and updates on schedule, expanding the usable corpus that models can read, cite, and act upon.
Both stream types will be available starting Q2 2026 as add-on annual licenses.
Vibe Coder: 100K stories/month for $5-$15/mo, perfect for hobbyists and early prototyping. Includes Single Stream reasoning.
Pro Developer: 1M stories/month for $20-$40/mo, ideal for serious developers and small-scale production workloads. Includes Single + Cross-Stream reasoning.
Enterprise: 10M stories/month for $200-$400/mo, built for production systems operating at scale. Includes Single + Cross-Stream reasoning.
All plans include Discovery streams (all-included). Narrative and Research streams require separate add-on licenses, coming Q2 2026. You can upgrade or downgrade at any time — changes take effect at your next billing cycle.
Web scraping gives you raw HTML. Building custom pipelines means maintaining parsers, handling changes, dealing with rate limits, and ensuring legal compliance. Synorb gives you structured, machine-ready intelligence with consistent formatting, verified sources, and fair-use compliance built in.
More importantly, we're not just aggregating data — we're synthesizing it. Each story is crafted for AI consumption: extracting entities, relationships, and insights that would require significant NLP infrastructure to replicate. You get signal, not noise.
Think of it this way: you could build your own database from scratch, or you could use PostgreSQL. You could scrape the web, or you could use Synorb. We're infrastructure, not a workaround.
Start with the free tier — 1,000 stories per month with no credit card required. Just enter your email above to receive access credentials instantly.
Test your credentials using our MCP Playground or API Playground. When you're ready to scale, signup for a paid plan.
Give your AI the context it deserves.