Signal8 MCP server / Dilution
Dilution tools for Claude, ChatGPT and Cursor
Seven MCP tools that give your AI assistant Signal8's small-cap dilution analysis: warrants, convertible notes, preferred stock, shelf registrations, ATM programs, equity lines, baby shelf capacity and cash runway. Every figure is cited to the SEC filing it was read from.
Ask in plain English
Once the server is connected, your assistant picks the right tools itself. Questions like these work as written:
- Run a dilution risk analysis on $TICKER.
- What warrants does $TICKER have outstanding, and at what exercise prices?
- How much can $TICKER raise off its shelf right now under the baby shelf rule?
- Is $TICKER's ATM still active, and how much capacity is left on it?
- How many months of cash does $TICKER have at its current burn?
- How did $TICKER trade after its last three offerings?
Or run the built-in explain_dilution prompt, which pulls the coverage, risk score, snapshot and instrument rows and explains each figure in plain English using the Signal8 Academy.
The seven tools
| Tool | What it answers |
|---|---|
get_dilution_coverageREST endpoint | Has Signal8 analysed this company, how fresh is the analysis, and if not, why not. Call it first. |
get_dilution_riskREST endpoint | The 0-100 Dilution Pressure Score and its seven components: shelf capacity, ATM and equity-line capacity, warrants in the money, convertible proximity, cash burn, historical dilution and toxic financing. |
get_dilution_snapshotREST endpoint | Potential new shares as a percentage of shares outstanding, per-family totals, cash position, monthly burn, runway and float. |
get_dilution_instrumentsREST endpoint | The rows for one instrument family (warrants, convertibles, preferred, shelfs, ATMs, equity lines, S-1 offerings, equity plans, exchangeables or recent offerings), each figure with the filing and sentence it was read from. |
get_baby_shelf_capacityREST endpoint | How much the company can raise off its S-3 shelf right now under the I.B.6 one-third-of-public-float cap, after the last twelve months of takedowns. |
get_dilution_performanceREST endpoint | How the stock traded after each of the company's past offerings. History, not a forecast. |
get_dilution_historyREST endpoint | Shares outstanding and cash over time from XBRL filings, with offerings overlaid so a raise lines up against the share count. |
Not measured is not zero
Dilution data is easy to misread in the reassuring direction. When a figure could not be measured or verified, the tools say so rather than returning a zero, and they tell your AI that a missing warrant count is not the same as no warrants. Totals built on withheld rows are flagged as lower bounds, scores report how much of the score could be measured, and every response carries the date of the filings it was read from.
New to dilution?
The free Signal8 Academy explains every instrument these tools report on. Your AI can read the same guides through the search_learn_guides tool, no API key needed.
Questions
- Which AI clients can use the Signal8 dilution tools?
- Claude, ChatGPT, Cursor, VS Code and any MCP client over Streamable HTTP. Connect the Signal8 MCP server once and the seven dilution tools are available alongside the rest of the Signal8 tools.
- Which plan includes the dilution tools?
- Signal8 Pro, at $59/mo or $588/yr. Pro also includes MCP and API access to every other Signal8 dataset.
- Which companies are covered?
- Small-cap US-listed issuers, where dilution moves the stock. Coverage is partial and growing, so get_dilution_coverage tells you whether a company has been analysed before you ask for the full picture.
- Where do the numbers come from?
- SEC filings. Every figure carries the accession number, the filing date and the quoted sentence it was read from, so you can check it against the original document.
- What does it mean when a figure is missing?
- It means Signal8 did not measure it. It never means the company has no warrants, no shelf or no dilution, and the tool descriptions tell your AI the same thing.