Data Connectors

Connect your data.
AI finds the rest.

50+ connectors spanning APIs, databases, warehouses, and observability platforms. Describe what you need — auto-discovery matches your query to the right source, authenticates, and starts streaming data in seconds.

Databases to dashboards in one sentence. No per-host pricing. No vendor lock-in.

Auto-Discovery

How auto-discovery works.

Three steps from natural language query to live data stream.

01

Describe

Tell the system what data you need. “Show me Bitcoin price vs S&P 500 over the last year.” The intent analyzer identifies entities, metrics, and time ranges.

Natural language input only — no API configuration required.

02

Discover

The API hunter agent searches 50+ providers across 35+ data domains. It scores each by reliability, rate limits, and data freshness — then selects the best match.

Intelligent ranking based on real-time performance metrics.

03

Connect

Credentials are requested once and stored per-user with encryption. The connector handles authentication, pagination, rate limiting, and schema mapping automatically.

Secure credential isolation with automatic retry logic.

Connector Catalog

Connector catalog.

35+ data domains. 50+ pre-built connectors. New providers added weekly.

External APIs

Data Domain

Finance

Stock prices, forex, economic indicators. Alpha Vantage, Frankfurter, FRED integrated.

Data Domain

Crypto

Token prices, DeFi protocols, on-chain data. CoinGecko, CoinCap, DeFi Llama support.

Data Domain

Weather

Forecasts, historical climate, air quality. Open-Meteo, OpenWeatherMap built-in.

Data Domain

Sports

Scores, stats, schedules. NHL API, BallDontLie for basketball and more.

Data Domain

Health

Disease tracking, public health data. Disease.sh for global health metrics.

Data Domain

Space

Launches, missions, earth observation. SpaceX API for real-time mission data.

Data Domain

Government

Census, economic, geological data. FRED, USGS, REST Countries available.

Data Domain

Developer

GitHub repos, API metrics, service status. GitHub API for repo analytics.

Databases & Warehouses

Data Domain

PostgreSQL / MySQL

Connect to your transactional databases. AI agents write optimized read-only queries, handle connection pooling, and respect row-level security.

Data Domain

ClickHouse / TimescaleDB

Time-series and analytical workloads. Sub-second aggregations over billions of rows with automatic query optimization.

Data Domain

Snowflake / BigQuery

Cloud warehouse native. Zero-ETL ingestion from your existing data pipelines. Federated queries across accounts and projects.

Data Domain

Databricks / Redshift

Lakehouse and columnar storage. Connect to your existing compute layer — no data movement required.

Observability & Metrics

Data Domain

Prometheus / OpenTelemetry

Ingest metrics, traces, and logs through open standards. No proprietary agents. No per-host pricing surprises.

Data Domain

CloudWatch / Cloud Monitoring

Pull AWS and GCP operational metrics directly. Infrastructure health alongside business KPIs in one view.

Data Domain

Custom Metrics Streams

Push any metric via HTTP, WebSocket, or MQTT. Schema-on-read means no upfront configuration.

AI Infrastructure

Intelligent routing across AI models.

Every query routes to the optimal AI model. Best answer. Lowest cost. Zero vendor lock-in.

Adaptive model selection means simple data lookups use fast, efficient models. Complex multi-source analysis uses frontier reasoning models. You get the best answer at the lowest cost.

Automatic failover ensures if a model provider experiences downtime, requests reroute instantly. Multiple redundant providers ensure 99.9% availability.

Prompt caching means repeated queries hit a semantic cache — 90% cost reduction and sub-50ms responses for cached content.

Zero vendor lock-in as the infrastructure adapts as new AI models ship. No code changes required when better models become available.

AI Routing Live status
Model A
Active — handling query
Model B
Standby — ready for failover
Model C
Standby
Cache hit rate 94.2%
Avg latency 23ms (cached) / 1.2s (fresh)
Cost saved 73% vs single-model

Authentication and security.

Every authentication pattern supported. Credentials isolated per user. Zero exposure.

Six auth types

API Key, Bearer Token, OAuth 2.0, Basic Auth, Query Parameter, Custom Headers. Every major authentication pattern supported out of the box.

Per-user credential isolation

Your API keys never leave your org boundary. Credentials stored with encryption, scoped per user, rotatable without downtime.

Rate limit management

The connector layer respects provider rate limits automatically. Requests queue, retry with exponential backoff, and distribute across time windows.

Data Processing

Schema intelligence.

Automatic field detection. Smart flattening. Type inference. Chart-ready data in seconds.

Automatic field detection means the data parser inspects every API response. It identifies time fields, numeric metrics, categorical labels, and nested structures without configuration.

Smart flattening transforms deeply nested JSON into chart-ready flat arrays. Reference fields (URLs, IDs) are detected and can be expanded.

Type inference auto-detects dates in any format (ISO 8601, Unix timestamps, human-readable). Currency fields, percentages, and units are recognized and formatted.

Field Analysis schema_parse
timestamp Time axis
price Numeric (USD)
volume Numeric (compact)
symbol Category
metadata auto-flatten
5 fields detected 3 numeric · 1 time · 1 nested

Unified Query

One query across everything.

Stop switching between twelve tabs to understand one incident. Ask one question — get answers from every data source simultaneously.

Unified data layer means your database metrics, application traces, business KPIs, and external market data — queried with one sentence. No separate tools for logs versus metrics versus traces.

AI that investigates surfaces anomalies before you ask. Traditional monitoring shows you charts and waits for you to interpret them. GlacierHub agents actively correlate signals across sources and explain root causes in plain language.

Predictable pricing with no per-host fees. No per-metric surcharges. No surprise bills when your infrastructure scales. One price for all your data, all your queries, all your users.

Open standards, zero lock-in as the platform is built on OpenTelemetry, Prometheus-compatible, and standard SQL. Your data stays portable. Switch or extend without rewriting instrumentation.

Unified Query Interface Live query
“Show me API latency spikes correlated with database load over the last 6 hours”
Prometheus
PostgreSQL
CloudWatch
App Traces
Correlated Result
Found 3 latency spikes (avg +240ms) between 14:30-15:15. Database CPU peaked at 87% during same window. Root cause: inefficient query on orders table (full table scan).
Query time 1.2s (4 sources)

By the Numbers

Performance stats.

Production metrics from organizations streaming live data.

50+

Pre-built connectors across APIs, databases, and warehouses

35+

Data domains from APIs to observability

6

Every pattern supported

< 50ms

Cached response latency

94%

Cache hit rate in production

99.9%

Uptime via automatic failover

Get Started

Start connecting data in seconds

50+ connectors. AI-powered discovery. Zero configuration. Get early access.

Get Early Access