summary: quick answers to the most common edgeful API questions: whether you need Claude, which AI tools work with the MCP connector, what data each plan gets, whether it costs extra, rate limits, and how to start.

no. the edgeful API is plain HTTP. you can call it with curl, Python, or JavaScript, with no AI tool involved at all.

Claude Code (or OpenAI's Codex app) is just an optional no-code convenience: it can read the docs and run requests on your machine for you. the regular ChatGPT website can't do that, so if you prefer the ChatGPT route you'd need the Codex app. none of this is required to use the API.

and if you just want to ask questions in a chat, the MCP connector does that with no key at all. which tools it works with is right below.

the connector works with these tools:

if your tool isn't listed, it needs to let you add a custom remote MCP server (Streamable HTTP) at https://api.edgeful.com/mcp, either with sign-in or with an API key through mcp-remote. if it can't add a custom MCP server, the connector won't work there.

Perplexity and TradingView's AI copilot also work with the connector, since both support custom MCP connectors. there's no dedicated walkthrough for them yet, but the setup approach is the same as for Claude or ChatGPT.

whichever tool you use, the connector reaches the same report catalog as the API, plus the discovery scan, with the same plan limits. setup steps for each tool are in connecting edgeful to Claude, ChatGPT, and Grok with MCP.

it depends on your plan. pro and all access get identical data through the API. the difference between those two is the algo suite, not the data.

what comes back is report output: the same numbers you'd see on the report page. on pro and all access, the response also carries the per-day rows behind those numbers in a detailed field. on essential, detailed is left out of the response, so you get the summary numbers only.

live data endpoints (what's in play + screener) are available on the pro and all access tiers. essential API access is historical report calculations only. the MCP connector uses the same plan limits as the API.

no. raw intraday OHLC bars, like 1-minute candles, can't be pulled through the API or the MCP connector.

what you get instead is report output: the summary stats you'd see on the report page, plus the per-day rows behind them in the detailed field on pro and all access.

no. API access is included with every plan and available right away. there's nothing to wait for.

no. API access is included on every plan at no extra cost — there's no separate "API tier" to buy. the $99/mo you might have seen is the pro plan price, not an API fee. what changes between plans is depth: which reports, tickers, and how much lookback the API exposes (essential 6mo / pro and all access 8+ years, no rolling cutoff), not whether you get the API at all.

they're the same across all tiers: 120 requests per 60-second window (sustained), a short burst allowance of 30 requests per 10 seconds, and a ceiling of 1,500 requests per hour. the limits apply per key, and your account has one active key at a time, so in practice that's your account's limit.

go over and the API returns a 429. caching responses for the trading day is the simplest way to stay well under the limit. full detail in rate limits and tier differences.

open the API dashboard, go to API Keys, and click generate API key. name it, then copy the value: it starts with ef_live_ and is shown only once, so store it somewhere safe. the base URL is api.edgeful.com, and you authenticate with a Bearer token in the Authorization header.

never written code before? the VS Code + Claude Code walkthrough gets you pulling data in about 10 minutes without writing a line.

the API is organized by report slug, not by metric name — there's no /probabilities or /stats endpoint. you request a report by its slug (like gap-fill-standard) against the base URL api.edgeful.com, and v1 is served at the root (no /v1 in the path). the settings and custom report routes are the exception: they do use /v1 (for example /v1/settings), see manage your settings with the API. the full catalog of valid slugs lives in the docs. if a call returns a 404, check the slug first — see API errors and troubleshooting.

the live what's in play stream (GET /live/{market_type}/wip) is a pro and all access feature. on essential, it returns a 403 live_data_not_allowed.

the stream carries what's in play data, so the reports it covers are the ones in the what's in play report selector. the full list is in selecting reports on what's in play.

full parameters are in the live WIP stream docs.