Agents
Context for your agents to take the best action
Trading agents are only as good as what they know the moment they act. Elfa gives yours live market, news and social context before every order, so it never trades into a story it hasn’t read yet.
Enrich with the latest context on events that matter
Fundamentals move markets: earnings, token unlocks, protocol upgrades and macro releases. Elfa catches these events as they break, tracks how fast they spread and maps them to the assets they move, so your agent reacts first.
Disambiguate between similar assets
WETH is ETH, and one ticker can point to five different tokens. Elfa maps every ticker, wrapper and nickname to the right asset, so your agent reads the whole conversation about what it trades, and none of the noise.
“It’s definitely made it easier to understand how calls performed after the fact. It’s also made it much faster to find callers and calls worth looking at, and it’s much less intensive than scanning X with AI tools. Someone was reporting other traders’ options flow as their own calls”
Event enrichment by the numbers.
Accounts tracked, ranked by signal quality
- 4.2M
Posts ingested today across every source
- 48M
Relevant mentions extracted today
- 6M
Spam and bot posts filtered before scoring
- 95%
Powering 3,000+ builders and product teams. From ambitious startups to major enterprises.
Questions and Answers
What data sources does Elfa use?
Posts from mainly X and Telegram and emerging social platforms, news, SEC filings, prediction markets like Kalshi and Polymarket, and market data including prices, funding and liquidations from exchanges Binance, Bybit and Hyperliquid. Every event links back to the posts behind it.
Which entities does Elfa cover?
Equities, tokens, projects, people and organisations, across traditional markets and crypto. Resolve any of them from a ticker, name, X handle, contract address, CoinGecko id, FIGI or CIK, and you get one id to use everywhere.
How quickly do events reach the API?
Elfa writes events up as the first posts arrive and keeps updating them as the story spreads, so the sources behind an event grow in real time. Each event records when we first saw it, so your agent can get the most fresh context.
Can I get push alerts, or do I have to poll?
Both. Query the REST API on your own schedule, or set up Auto: describe a condition and Elfa evaluates it continuously, then delivers to a webhook, a Telegram bot or an SSE stream the moment it fires.
How do you tell a real event from a rumour or a duplicate?
Only events judged worth acting on are served, never spam. Repeat reports are merged into one event, follow-ups are threaded into a storyline that marks when a claim is confirmed, and only accounts in good standing count as sources.
How does Elfa plug into my agent?
Through the REST API, the TypeScript and Python SDKs, or the MCP server for Claude, ChatGPT, Cursor and VS Code. Agents with a shell, like Claude Code or Codex, can install the Elfa skill instead.
Can my agent ask questions in natural language?
Yes. Chat answers market questions in plain language with live context, and can stream its reply. Builder Chat also turns a plain-language request into an Auto condition, ready to validate and switch on.
How does pricing work, and can I try it for free?
Start on the Free plan with 1,000 credits a month at 60 requests per minute. Grow is $870 a month with 120,000 credits, pay-as-you-go is $0.0145 per credit with a $20 minimum top-up, and Enterprise is custom.