Labeled data, ready for models and agents.
Building on news and filings usually starts with scrapers, parsers and cleanup. Sitch delivers deduplicated records in one versioned schema, already labeled with tickers, sentiment, categories and places.
Load it into your warehouse, sync it incrementally for retrieval and analytics, or give your agents direct, read-only access through an MCP server.
The hard part today
- Scrapers and parsers break and need constant upkeep
- Every source arrives in a different shape
- Labels for entities and sentiment have to be built in-house
- Agents need reliable, sourced data to ground their answers
With Sitch
- Warehouse sync
- Retrieval corpora
- Training and evaluation sets
- Agent tools
What you can build
Built around your workflow.
- Warehouse sync
- Records loaded into your database or warehouse, with incremental updates
- Retrieval corpora
- Clean, labeled text with sources for RAG and search
- Training and evaluation sets
- Labeled records by ticker, category and sentiment, with history
- Agent tools
- MCP tools so agents can look up news, filings and trades with sources
Example records
The data behind it.
Real records from the feed. Links open the original source.
- NewsCNBC's The China Connection newsletter: Starbucks enters Xinjiang in calculated geopolitical riskCNBC$SBUXneutralgeopoliticscorporate
- XProperty buyers are demanding sellers renegotiate terms because of higher mortgage rates, signaling broader market distress.@wsjnegativemacroeconomy
- SEC filingChevron (CVX) filed an 8-K current report on a material eventSEC EDGAR$CVX8-K
- Insider · Form 4Duffy Matthew Paul, President & Chief Risk Officer, sold NP shares ($34.2K)SEC Form 4$NPnegativeinsider activity
- CongressMichael Rulli (R-OH) sold MSFT, $1,001 - $15,000House disclosure$MSFTcongressional trade
Data used
Built on these sources.
Delivered as: Real-time stream · REST API with backfill · Webhooks · Your database or warehouse · Slack, Teams or Discord alerts · Scheduled CSV/JSON exports · MCP for AI assistants · Custom dashboards
FAQ
Data & AI teams, answered.
Is the schema stable?
Yes. Records follow a versioned schema; new fields are added without breaking existing ones.
How do incremental updates work?
A live stream delivers new records, and cursor-based recovery lets you resume from your last checkpoint without gaps.
Can our agents use the data directly?
Yes, through the MCP server's read-only tools, or through the REST API.
Keep exploring
More ways to use Sitch.
Talk to us
Tell us what you're watching.
Every plan is scoped to your sources, filters and delivery, so there's no one-size price list. Send us a note and we'll come back with a sample of your feed.
- 01Tell us what you needSources, companies, topics and regions, and where the data should go.
- 02We scope it with youFields, filters and delivery, plus a sample of your feed.
- 03You go liveAPI keys, a feed into your systems or a custom dashboard.