Frontier Markets

37 million structured financial records. 1,729 US microcaps. 29.1M financial facts. 837K daily market records. 111K insider transactions. 118K company linked news events. 74K short interest observations. 24K government contract awards. Underneath it sits roughly 39 billion characters of source data and more than 6 billion word like tokens. This is what Frontier agents have to work with. Not another AI wrapper generating summaries from a few headlines. A financial intelligence layer built to connect fundamentals, price action, insider activity, contracts, attention, positioning and catalysts across thousands of emerging assets. The goal is to find the change before it becomes obvious. Screen millions of signals. Surface asymmetric setups. Show the evidence. Monitor them 24/7. And as these markets move onchain, connect intelligence directly to execution and liquidity. This is Frontier. 💎 Access is now open terminal.frontiermarkets.ai

The last few days we’ve been busy building something pretty new for this space. We’re using Jev to push toward much more accurate research reports while keeping the cost low enough for agents to run them constantly, not just once in a while. The interesting part is what happens when deep research becomes cheap enough to run across hundreds or thousands of companies at a time. Better reports, more verification, more sources checked, and a price that actually makes sense for agentic workflows. Still putting the pieces together, but we think this opens up a very different way for agents to do financial research at scale.

Arkham style exploration for stocks is now live on Frontier. Frontier Graph, a full page workspace for exploring the connections between companies, reporting owners, government agencies, programs, filings, news and themes. Start with a stock and expand its network. Rich entity cards surface financial metrics, disclosure context, dates and evidence counts. Click a connection to inspect its supporting records and open the original source. Select multiple companies to discover shared connections or trace a path between them. Filter by date, relationship type and reported value, then use the timeline to explore the evidence. Build a research tray, compare companies, save and share your exploration, or send your selection into the $100k historical portfolio simulator. Open Relationship graph from any stock’s details page and start exploring. terminal.frontiermarkets.ai/…

A major upgrade is taking shape. Built on 37M+ structured research records, we’re building an interactive research graph that lets you start with one company and follow its connections across insiders, filings, government awards, programs, news, themes, and more. Rich entity cards, expandable networks, evidence behind every connection. Here’s a first look.

Think Arkham-style exploration, but for public market intelligence. We’re building a graph layer on top of Frontier’s 37M+ records so you can move through the connections between companies, insiders, filings, contracts, news, themes, counterparties and catalysts instead of looking at everything in isolation. Start from one company, follow the graph outward, open the underlying evidence, and see relationships that are easy to miss in a normal screener. The bigger part is making this agentic too. Agents should be able to traverse the graph, connect signals across sources, and surface relationships worth investigating without you manually stitching everything together. Soon.

Just over the last few days, we’ve injected 808,651 new records and source documents into Frontier’s research loop: • 676,429 current SEC facts • 99,930 market records • 10,391 insider transactions • 3,333 company-linked news events • 3,819 short-interest observations • 13,375 government contract awards • 1,374 verified issuer documents This data isn’t collected just to sit in a warehouse. Every record is normalized, linked to the correct company, validated, scored, and published back into the terminal. As new filings, transactions, awards, price movements, and company updates arrive, the research and rankings evolve with them. The pipeline now runs across 1,725 US microcaps. Collect → validate → score → publish → repeat. The dataset compounds. The loop keeps running. terminal.frontiermarkets.ai

The $100k Portfolio Simulator is now live on the terminal. Pick a historical start date, choose up to 20 stocks, and decide how to allocate a virtual $100,000. Leave part in cash or use equal weights to build your portfolio. See how it would have performed with an interactive portfolio view, including its value over time, profit and loss, total return, maximum drawdown, and each stock’s contribution. Go from researching individual companies to exploring how they would have worked together in a portfolio. Open the terminal and click $100k Simulator in the top bar. terminal.frontiermarkets.ai

You can literally connect your AI apps to 37 million structured financial records through Frontier. Financials, prices, filings, insider activity, contracts, news, sentiment and more, all accessible through the same intelligence layer. And once our dedicated agent is live, you’ll be able to take that a step further with automated trading through Frontier Private Exchange.

Introducing the Frontier Markets MCP server. Connect Codex, Claude, Hermes, or another MCP compatible agent to research underfollowed companies using financials, prices, SEC filings, insider transactions, and source-linked evidence. In this video, we asked Codex to find a small company winning meaningful new business that hasn’t yet shown up in revenue, then investigate whether it’s an opportunity or a trap. The workflow goes beyond finding a ticker. Follow the evidence, examine the risks, and open the sources and terminal links to check the thesis yourself. MCP address: terminal.frontiermarkets.ai/…

Next week starts with major updates. Tomorrow, we’re releasing the Frontier MCP server. You’ll be able to connect Frontier’s intelligence data directly to Claude Code, Codex, and other MCP-compatible tools, with company financials, price history, filings, catalysts, and the evidence behind potential opportunities available inside the tools you already use. Then later in the week, we’re aiming to release our dedicated agentic bot. The idea goes much further than asking questions. Give it a company, theme, thesis, or goal and let it investigate on its own, follow the evidence across multiple sources, compare opportunities, track catalysts, recheck assumptions as new data comes in, and keep monitoring until something actually changes. More on that later.

3M $FRONT tokens bought back today. That’s the largest single buyback we’ve done so far. We’re keeping the process simple. Exchange activity generates fees, part of those fees goes back into the market, and we post the numbers publicly as it happens. We’ll keep doing this as usage of the Private Exchange grows. TX: robinhoodchain.blockscout.co…

What “agentic” means for Frontier is that the system won’t just answer a question and stop there. An agent can keep working toward a goal over time. It can watch for changes across fundamentals, filings, insider activity, contracts, news, sentiment and price action, decide when something is actually worth surfacing, compare it against your strategy, and keep tracking the thesis after that. It can also react to conditions you define. Alert me if this company wins a new contract. Find names where growth improves before price catches up. Watch this position and tell me if the thesis weakens. Prepare a trade if the setup reaches my rules. The important part is the loop. Observe, reason, act, keep monitoring. That’s the kind of agentic system we’re building into Frontier.

Early next week Frontier gets agentic. We’re launching our MCP server alongside the first version of the Frontier agent, opening up a much more direct way for agents to use the same intelligence, market data and execution rails behind the platform. The agents will be able to scan thousands of companies, follow fundamentals, insider activity, contracts, news, sentiment and price action, find setups that match a strategy, explain what changed, monitor a thesis 24/7, pull live quotes and prepare trades. You’ll be able to give it a goal, set the boundaries and let agents do the work around it. The MCP server also means other agents and agentic apps can start using Frontier as part of their own workflows. More on both soon.

Over the last 7 days alone, our website and terminals handled more than 127,000 requests through Cloudflare. That includes 74,000+ uncached requests hitting the backend directly. We’ve been scaling infrastructure around this growth, but seeing this level of activity this early is a strong signal for us. More people are using the terminals, testing the exchange, pulling data, and interacting with the product every day. We’re building for a much bigger load from here. Good problem to have.

$ALLT IS TURNING CYBERSECURITY GROWTH INTO RECURRING CASH FLOW Allot’s Q2 revenue grew 15% YoY to $27.7M. But the revenue mix is the more interesting part. SECaaS revenue: $9.4M, +47% YoY SECaaS ARR: $36.1M, +44% YoY SECaaS share of revenue: 34% Recurring revenue: 67% of total Deferred revenue also increased both sequentially and YoY, adding visibility into the rest of 2026 and 2027. And the operating leverage is already showing: → GAAP operating income: $1.1M vs. a $0.4M loss YoY → GAAP net income: $2.6M vs. a $1.7M loss YoY → Q2 operating cash flow: $8.5M → H1 operating cash flow: $19.0M → Cash + deposits + investments: $107M → Debt-free balance sheet Management raised 2026 revenue guidance to $115M-$118M and continues to expect SECaaS growth of at least 40%, alongside further profitability improvement. The next part of the thesis may be capital allocation. Allot authorized a $40M share repurchase while continuing to generate cash and scale its recurring cybersecurity business. The setup is shifting from a legacy telecom vendor story toward a profitable, recurring-revenue cybersecurity business with a strong balance sheet. Check the evidence in our terminal: terminal.frontiermarkets.ai/… Powered by $FRONT

Another $FRONT buyback is complete. Today, 1.4M tokens were purchased using fees generated by Frontier Private Exchange over the last 24 hours. This follows yesterday’s first buyback and continues the mechanism of routing exchange fees back into $FRONT as trading activity grows. Buyback tx: robinhoodchain.blockscout.co…

We are now officially live as an organization on the npm registry. @frontiermarkets/privacy-sdk is now published under the Frontier Markets namespace and available publicly through npm. This gives Frontier an official distribution layer outside the terminal itself and makes the private exchange infrastructure accessible as a product of its own. Install it in one command: "npm install @frontiermarkets/privacy-sdk" Frontier Markets on npm: npmjs.com/package/@frontierm…