Why we’re building
Ultralayer.

We’ve spent years building systems that use AI for financial analysis. A few things became clear:

1. Context efficiency is key.
   • Cost: tokens are expensive, and agents amplify this problem. Packing the same information into fewer tokens = lower cost.
   • Quality: even though this is continuously getting better, reasoning quality still degrades as context increases.
   • Hallucination: as context grows, the risk of hallucination increases. This is especially important for financial analysis.
2. Market data APIs weren’t built for AI agents.
   • They were built for quant work and dashboards.
   • Market data is still necessary for AI financial analysis. However, the legacy APIs miss the qualitative side: market context, evidence, explanation behind the numbers, and so on. This is the type of information that agents are optimized to work with.
3. Anyone building similar systems would be rebuilding the same tools.
   • We saw an opportunity to abstract these primitives into an API and serve them.

We monitor the market, take care of these needs, and serve it to you at a lower cost than you could build it yourself.

Our goal is to solve valuation and improve capital allocation in the public equity markets.

We expect AI to become more capable of making informed financial decisions than the average person. We want to build the systems that support this, and make them accessible to the public.

These are the design principles that we’re committed to:

• Sources and evidence: all claims include sources. Quoted evidence is present where applicable. We check that sources exist, that quoted text matches the retrieved material, and that it supports the claim.
• Context efficiency: two examples; the MCP tools + schemas, and a response detail parameter (essential / standard / full) so your AI only retrieves the context it needs.
• Our own datasets: all context we serve is generated by us. The only time it isn’t is when we’re quoting source material.
• Backtesting: we differentiate between when a source was knowable and when we ingested it. This provides point-in-time safety and enables backtesting.
• Coverage expansion: there’s a general focus on the US market today, and we’re built to expand coverage as the system matures. Some tools already have global coverage.
• Access and affordability: we’re designing things to be cost-effective and accessible to the retail public.

Our aim is institutional quality at pricing that works for individuals.

• Free trial credits when you sign up.
• Pay as you go, usage based pricing.
• No required subscription or minimum spend.
• The same underlying rates through the API, MCP, and plugins.
• No additional surcharge for alerts.

For context, $1 covers approximately:

• 330 Wire requests, returning 20 headlines each.
• 100 full development briefs, including stakeholders and evidence.
• Six stakeholder analyses, with assessed impacts and company comparisons.

We’re committed to making the best systems we can, and keeping them accessible.