Find setups using natural language.
Our semantic data layer maps complex trading patterns into simple queries, making your agent work smarter, faster, and cheaper.
Hundreds of indicators optimized for agents.
We transform raw market data using categorical bands, improving agent reasoning and significantly reducing token usage.
| Field | Value | Why it matters |
|---|---|---|
trend_ma_alignment | aligned_bullish | Shows whether short, intermediate, and long-term averages agree. |
trend_ma8_slope | steep_up | Captures the fastest moving-average velocity state. |
trend_ma20_slope | shallow_up | Converts the 20-period MA slope into a readable velocity state. |
trend_ma40_slope | flat | Tracks the stage-analysis average velocity on daily and weekly candles. |
trend_ma50_slope | shallow_up | Adds intermediate trend velocity used by many screeners. |
trend_ma100_slope | shallow_up | Shows mid-to-long-term trend acceleration or decay. |
trend_ma200_slope | flat | Anchors long-term trend velocity without chart parsing. |
trend_ma_compression_band | compressed | Shows whether the 8, 50, and 200 MAs are coiled or fanned out. |
trend_ma_crossover_event | bullish_trend_cross | Highlights the highest-priority MA cross on the current candle. |
trend_distance_ma8 | proximity_above | Captures very short-term extension or weakness. |
trend_distance_ma20 | extended_above | Useful for current swing context. |
trend_distance_ma40 | proximity_above | Connects daily context to longer-term stage analysis. |
trend_distance_ma50 | extended_above | Common intermediate trend reference. |
trend_distance_ma100 | proximity_above | Adds mid-range trend context without chart parsing. |
trend_distance_ma200 | extended_above | Anchors the agent to the long-term trend line. |
Ask your agent to compare candles.
Get a clean, structured diff that clearly highlights every technical movement. We handle the data behind the scenes to ensure you get answers faster.
Integrate into ChatGPT, Claude or Gemini in minutes.
Connect TickerDB through MCP and your LLM can call market data tools directly from the chat. No custom parsing, no indicator math, no extra data pipeline.
Add to any codebase in minutes.
Add our MCP server, SDKs or direct API calls to any project.
Loading example...
Loading staged response...
FAQ
What is TickerDB?
TickerDB is a categorical-first market intelligence database covering 10,000+ active US stocks, ETFs, and supported crypto pairs with 5+ years of history.
It packages market-state calculations into named bands your agent can work with directly. Responses include supporting metadata like durations, percentiles, and support/resistance context.
Query it via the plain HTTP API, MCP, or an SDK when you want derived market state and context without rebuilding that classification layer yourself.
Who is TickerDB for?
TickerDB is for anyone building agentic trading bots, market research tools, portfolio analysis workflows, or financial data products that need structured market context.
That includes developers, financial analysts, market researchers, financial advisors, quantitative researchers, and product teams building AI-assisted finance tools.
If your application needs an agent to reason about market state without rebuilding the data pipeline and classification layer, TickerDB is built for that.
How does search work?
Search gives you a database-like query layer over the TickerDB dataset.
You can combine filters across market state, fundamentals, volume, momentum, volatility, sector context, and asset metadata. For example: momentum_rsi_zone = oversold AND sector = Technology AND volume_ratio_band IN (high, extremely_high).
You can sort server-side, including by fields like market_cap, select only the fields you want back, and paginate through results.
It feels like querying your own market database, but TickerDB manages the pipeline, computes the derived fields, and keeps the published data checked and current.
How does TickerDB help agents use fewer tokens?
TickerDB gives agents pre-computed market context instead of raw candle data.
Rather than sending long OHLCV histories, fundamentals, and indicator calculations into the prompt, you can send compact fields like momentum_rsi_zone, trend_direction, volume_ratio_band, and support_distance_band.
That reduces the amount of data the model has to read while improving reasoning, because the important market state is already labeled, structured, and ready to compare.
You can also select only the fields your workflow needs, keeping each response focused on the decision your agent is trying to make.
What assets do you support?
10,000+ active assets across US stocks, ETFs, and supported crypto pairs.
Pass any ticker to /v1/summary/TICKER - if it returns data, it is covered. If not, you get a 404.
No separate assets list endpoint needed.
How often is the data updated?
Daily end-of-day data only. Stocks and ETFs usually refresh by approximately 00:30 America/New_York, while crypto usually refreshes by approximately 00:10 UTC, including weekends.
Every response includes "data_status": "eod" for the freshness mode and "as_of_date" for the exact session date represented by the snapshot.
No intraday data - TickerDB is end-of-day only. See Data Quality for the full freshness notes.
What are categorical bands?
Categorical bands are named buckets for computed market conditions. Instead of making your app interpret raw numbers, TickerDB gives you fields an agent can reason about directly.
For example, RSI can be exposed as "momentum_rsi_zone": "oversold", volume can be exposed as "volume_ratio_band": "high", and support distance can be exposed as "support_distance_band": "near".
The underlying numbers still matter, but the band gives your bot, screener, or research workflow a clean state label to filter, sort, compare, and explain.
How much historical data is available?
Currently we support 5+ years of data.
Historical access depends on your plan: Starter includes 30 days, Plus includes 2 years, and Pro includes the full history.
What type of architecture does TickerDB run on?
The core dataset is stored in TimescaleDB on high-availability colocation clusters running on bare-metal machines. Data is computed, validated, and published through a staged pipeline before it becomes available to the API.
Data in high demand is hot-cached on Cloudflare KV at the edge, so common reads stay fast without doing request-time market calculations.
Your application gets database-backed durability, edge-cached reads, without having to run your own infrastructure.