Introducing the SentiSense API: Financial Intelligence for Builders and AI Agents

If you have ever tried to build a financial dashboard, a trading copilot, or a research agent on top of public market data, you already know the problem: the data exists, but it is scattered across SEC filings, social platforms, news APIs, and a half-dozen vendors that each charge separately for one slice of the picture.

We have spent the last year stitching that picture together for our own product. Today we are exposing it as a clean, public API.

This post is the launch note. It covers what is in the API, what is free, what costs money, what the SDKs look like, and what an AI agent can do with it out of the box.

What you can do with the SentiSense API

The API is read-only. You will not place trades through it. You will not move money. What you will get is a single, consistent surface for the data and analysis we already use to power app.sentisense.ai.

Today's launch covers nine API groups:

1. Stocks (27 endpoints)

The core company-data surface. Quotes, profiles, sector and industry classification, historical price data, technical indicators, dividends, earnings calendar, fundamentals, and as of today, company KPIs: structured per-segment revenue and operating metrics extracted from 10-Q and press release data. Think iPhone Revenue, AWS Revenue, Cybertruck deliveries, Data Center Revenue, with full historical series.

GET /api/v1/stocks/AAPL/kpis
GET /api/v1/stocks/with-kpis
GET /api/v1/stocks/{ticker}/kpis/types

We have curated KPIs for the S&P 500 and an extended universe (around 500 tickers today, expanding into the long tail of US equities).

2. Institutional Flows (7 endpoints)

13F holdings, top institutional movers, ownership trends, and as of today, a per-institution detail endpoint that gives you a fund's full reported book and its largest position changes quarter over quarter. Bridgewater, Berkshire, Renaissance, Citadel, Two Sigma: anyone who files a 13F is in there.

GET /api/v1/institutional/institution/berkshire-hathaway
GET /api/v1/institutional/institution/bridgewater-associates

3. Politicians Trading (4 endpoints)

Congressional disclosed trades, by member, by ticker, and by recent activity. Free preview for unauthenticated callers; full history with PRO.

4. Insider Trading (3 endpoints)

SEC Form 4 filings: officer and director purchases and sales, with cluster detection so you can see when multiple insiders move at the same time.

5. Analyst Ratings (formal API as of today)

Wall Street ratings, price targets, estimates, and rating change history. Includes a consolidated /summary endpoint that gives you the current consensus, distribution, and recent revisions in a single call.

GET /api/v1/analyst/AAPL/summary
GET /api/v1/analyst/AAPL/estimates
GET /api/v1/analyst/AAPL/changes

6. Market Mood (formal API as of today)

Our proprietary 0 to 100 fear and greed composite, blending five sub-signals (Social Sentiment, Market Direction, Fear Gauge, Social Momentum, S&P 500 Trend). Free for all callers, no API key required, with full per-sector breakdowns and 6 months of daily history. We wrote about how this differs from CNN's Fear and Greed Index in an earlier post.

GET /api/v2/market-mood?days=180

7. AI Insights (6 endpoints)

This is the proprietary surface that does not exist anywhere else. Our knowledge graph plus our agentic pipeline produce daily insights tied to specific tickers: divergences between price and sentiment, news clusters worth attention, signal vs price moves, and so on. Today we shipped three new endpoints that make this surface usable from outside the SentiSense web app:

GET /api/v1/insights/range/{from}/{to}
GET /api/v1/insights/latest
GET /api/v1/insights/user

These let you pull a window of insights, the latest cross-ticker batch, or the personalized stream for a specific user.

8. Documents & News (7 endpoints)

The news feed that powers our sentiment and clustering. Per-ticker news, sentiment-annotated documents, news clusters, and a sitemap. The API returns derived data (sentiment, reliability, source, URL, timestamp) but intentionally does NOT return article headlines or article body content. That is a publisher-copyright constraint, not a missing feature. If you need the headline, follow the URL.

9. Stories

Long-form, agentically composed narratives that synthesize a multi-day news arc. Useful as input to summarization agents, or as a "what is the big story this week" surface for a dashboard.

What is free, what costs money

We have four access tiers, and as of today, a fifth one:

Public: no auth required at all (Market Mood, sitemap, terms).
Public preview: unauthenticated callers get a preview snapshot, with a previewReason flag in the response telling them what is gated. PRO callers get the full payload.
Discovery (key required, no quota cost): new today. Endpoints like /with-kpis and /{ticker}/kpis/types require an API key (so we can track usage per identity), but do not consume your monthly quota. These are metadata about data, not the data itself.
Quota-gated: the heavy data endpoints. Every call counts against your monthly quota. FREE tier ships with 1,000 requests per month, which is enough to build and run a personal dashboard or a small research project.
PRO only: full historical depth, full institutional and politician trade history, full insights, no preview wrapper. $15/month, unlimited within reasonable rate limits.

Per-minute rate limiting applies to all tiers and bounds abuse.

SDKs: Python and Node, both shipped today

You can use the API with curl, requests, or fetch. Or you can install one of the SDKs.

Python (PyPI)

pip install sentisense

from sentisense import SentiSenseClient

client = SentiSenseClient(api_key="ss_live_YOUR_KEY")

# Latest insights across all tickers we tracked today
latest = client.get_latest_insights()
for insight in latest:
    print(insight.ticker, insight.title)

# Apple's revenue mix by segment
kpis = client.get_company_kpis("AAPL")
for series in kpis.data.kpis:
    print(series.name, series.values[-1].value)

# What the market mood is right now
mood = client.get_market_mood(days=30)
print(mood["market"]["currentScore"], mood["market"]["phase"])

Version 0.12.0 ships with typed dataclasses for the KPI surface, so your IDE actually knows what kpis.data.kpis[0].values[0].value is.

Node (npm)

npm install sentisense

import SentiSense from "sentisense";

const client = new SentiSense({ apiKey: process.env.SENTISENSE_API_KEY });

const summary = await client.analyst.getSummary("NVDA");
console.log(summary.consensus, summary.priceTarget);

const top = await client.institutional.getTopInstitutions();
const berkshire = await client.institutional.getInstitutionDetail("berkshire-hathaway");

Both SDKs are MIT-licensed, source on the third-party repos.

The AI agent skill

This is the part we are most excited about, because nobody else has shipped it.

The SentiSense skill is a single markdown file at sentisense.ai/skill.md that any modern AI coding agent (Claude Code, Cursor, OpenClaw, ClaWHub-compatible runtimes) can ingest in one step. Once installed, the agent knows:

Every endpoint, parameter, and response shape
How authentication works
Which tier each endpoint belongs to
Which endpoints consume quota and which do not
Which idioms to prefer for common research tasks

Install via ClawHub:

npx clawhub@latest install sentisense

Or just point your agent at the URL. We treat skill.md as a first-class API contract: when we ship endpoints, the skill ships with them on the same day.

This is what we mean when we say SentiSense is built for AI agents. It is not a chat UI bolted onto a database. It is a data API plus a published skill so an agent can be a competent researcher on day one.

What you can build in an afternoon

A few things we have built or seen built on the API in the last two weeks:

A personal earnings dashboard that pulls upcoming earnings dates, the analyst consensus, and the social sentiment trend for a watchlist, then renders one card per ticker.
A 13F drift tracker that diffs Berkshire and Bridgewater positions quarter over quarter and posts the deltas to a Slack channel.
A politician copy-trade alert that watches recent congressional trades, filters for purchases over $50k, and emails the result.
A macro mood widget for a broader fintech site, just embedding the Market Mood score and weekly change.
A research agent that, given a ticker, pulls the latest insights, the news cluster summary, the analyst summary, the top institutional movers, and the KPI series, and writes a one-page brief.

If you build something, send it to hello@sentisense.ai. We will likely feature it.

Honest caveats

A few things to keep in mind:

Quotes and prices are real-time. Sentiment, insights, and news clusters are batch. Insights regenerate daily. The generatedAt timestamp on every insight tells you exactly how fresh the snapshot is. If you need intraday signals, use the quote endpoints, not the insight ones.
Headlines are not in the API. Article URLs, sources, sentiment, and reliability are. Headlines stay with the publishers; follow the URL if you need them.
KPI coverage is near-complete on the S&P 500 plus extended universe today. The long tail of US tickers is on the roadmap. Use /with-kpis to see the current coverage list before you assume a ticker is in there.
This is informational data, not investment advice. Market Mood, insights, sentiment, and analyst summaries are tools for research, not signals to trade on. Always do your own work.

Get started

Create a free account at app.sentisense.ai/signup.
Generate an API key in Settings > Developer Console.
Install an SDK or read the API docs.
Drop the skill into your AI agent.
Build something.

The free tier is generous. The PRO tier is $15/month. The skill is free forever.

Welcome to SentiSense.

SentiSense provides market information, sentiment analysis, institutional flow data, and data intelligence for informational and educational purposes only. We do not provide investment advice, recommendations, or financial guidance. Users are solely responsible for their own investment decisions.