# StackSpend Documentation > StackSpend is a cloud and AI cost visibility platform for engineering teams. It connects to cloud providers (AWS, GCP, Azure), AI platforms (OpenAI, Anthropic, Cursor, Claude), data platforms (Snowflake, ClickHouse), and communications tools (Twilio, GitHub, Vercel) to provide centralized spend tracking, anomaly detection, budget alerts, and daily cost reports. ## Getting Started - [Quick start](https://docs.stackspend.app/getting-started): Connect your first provider and get costs visible in under 20 minutes. - [Connecting providers](https://docs.stackspend.app/connecting-providers): Overview of every supported provider with setup time, data latency, and historical data coverage. ## Features - [Dashboard](https://docs.stackspend.app/features/dashboard): At-a-glance spend overview with provider breakdown, cost trends, and forecasts. - [Data Explorer](https://docs.stackspend.app/features/data-explorer): Interactive cost analysis — filter by provider, service, category, project, user, or date range. Group, chart, and export. - [AI Explorer](https://docs.stackspend.app/features/ai-explorer): Cross-provider LLM usage lens — usage by model, provider, project, and user, measured in tokens and API-equivalent value, with input/output/cache token drill-downs. - [Model recommendations](https://docs.stackspend.app/features/model-recommendations): Benchmark-guarded suggestions for cheaper equal-or-better LLM alternatives, priced at your real token mix, with projected savings and a dismiss/implement lifecycle. - [Spend monitoring](https://docs.stackspend.app/features/monitoring): Traffic-light (green/amber/red) health view across all providers showing MTD, projected month-end, and budget status. - [Forecasting](https://docs.stackspend.app/features/forecasting): ML-powered month-end spend forecasts with confidence intervals and days-to-risk estimates. - [Inbox](https://docs.stackspend.app/features/inbox): Triage panel for alerts (anomalies and budget threshold breaches), setup tasks, and opportunities. Docked to the right side of the screen. - [Tasks](https://docs.stackspend.app/features/tasks): Company-wide table of open alerts, providers without budgets, unowned providers, and setup tasks. - [Anomaly detection](https://docs.stackspend.app/features/anomaly-detection): Automatic detection of spend spikes, drops, trend changes, and new services after every sync. - [Source control correlation](https://docs.stackspend.app/features/source-control): Links a cost anomaly to the pull requests and deployments that landed before it, ranks the likely change, and explains why with the actual diff. Investigation aid, never asserts causation. Business plan. - [Alerts & escalation](https://docs.stackspend.app/features/alerts): Alert routing, severity levels, lifecycle states, and configurable per-severity escalation SLAs. - [Assigning owners](https://docs.stackspend.app/features/owners): Route alerts to the right person by assigning a provider owner. Unresolved alerts escalate to admins. - [Budgets & alerts](https://docs.stackspend.app/features/budgets): Set monthly spend limits per provider with configurable threshold alerts at 50%, 80%, and 100%. - [Tagging](https://docs.stackspend.app/features/tagging): Custom labels with auto-tagging rules for cross-provider cost grouping. - [Daily report](https://docs.stackspend.app/features/daily-report): Morning cost digest via email and Slack with yesterday's spend, MTD, open alerts, and budget status. ## Provider Setup Guides - [OpenAI](https://docs.stackspend.app/providers/openai): API Key + Organization ID. Real-time. Full history. - [Anthropic](https://docs.stackspend.app/providers/anthropic): Admin API Key (not regular API key). Real-time. Full history. - [Claude (OTel)](https://docs.stackspend.app/providers/claude): OpenTelemetry-based estimated usage from Claude Code and agents. - [AWS](https://docs.stackspend.app/providers/aws): IAM user with Access Key ID + Secret Access Key. 24–48 hour latency. 90 days history. - [GCP](https://docs.stackspend.app/providers/gcp): BigQuery billing export + service account JSON key. 24–48 hour latency. No historical backfill. - [Azure](https://docs.stackspend.app/providers/azure): App registration (service principal) with Cost Management Reader role. Daily. 90 days history. - [Cursor](https://docs.stackspend.app/providers/cursor): Team API key (Admin scope), team admin access. Real-time. Full history. - [GitHub](https://docs.stackspend.app/providers/github): Classic PAT with admin:org scope. Daily. From setup date only. Copilot per-member token usage and model mix also flow to the AI Explorer when the org's "Copilot usage metrics" policy is enabled. - [Snowflake](https://docs.stackspend.app/providers/snowflake): ORGADMIN or ACCOUNTADMIN credentials. Daily. 90 days history. - [Vercel](https://docs.stackspend.app/providers/vercel): API token + Team ID. Daily. 90 days history. - [ClickHouse Cloud](https://docs.stackspend.app/providers/clickhouse): API Key ID + Secret. Daily. 90 days history. - [Databricks](https://docs.stackspend.app/providers/databricks): Service principal (Client ID + Secret) + SQL warehouse, reads system.billing tables. Daily. 90 days history. - [Elastic Cloud](https://docs.stackspend.app/providers/elastic): Organization API key with Billing admin role. Daily. 90 days history. - [Hugging Face](https://docs.stackspend.app/providers/huggingface): Fine-grained access token with billing scope. Monthly billing cycle. - [Grok (xAI)](https://docs.stackspend.app/providers/grok): xAI Management API key. Real-time. Full history. - [Twilio](https://docs.stackspend.app/providers/twilio): Account SID + Auth Token. Daily. 90 days history. - [Custom provider (FOCUS)](https://docs.stackspend.app/providers/custom): Any cost source not natively supported. Push-only — you send fully-priced rows in the StackSpend FOCUS Profile (a profile of FinOps FOCUS 1.2) via the ingest API, or upload a CSV and map your columns. CSV import works on any plan; the ingest API requires the Business plan. ## API & Developer - [API reference](https://docs.stackspend.app/api): REST API for cost line items, daily rollups, and anomaly alerts. Business plan required. - [OTel ingest](https://docs.stackspend.app/api/otel-ingest): OpenTelemetry endpoints for Claude and Claude Code agent usage. - [Webhooks](https://docs.stackspend.app/api/webhooks): HTTP POST notifications for anomaly.created events with HMAC signature verification. - [Claude Code (MCP)](https://docs.stackspend.app/developer/claude-code-mcp): Query StackSpend cost data from Claude Code via the Model Context Protocol. ## Account - [Team management](https://docs.stackspend.app/account/team): Invite teammates, assign Admin/Member roles, manage pending invites. - [Settings](https://docs.stackspend.app/account/settings): Organization name, timezone, preferred currency, forecasting thresholds, notification preferences. - [Billing & plans](https://docs.stackspend.app/account/billing): Team, Business, and Enterprise plans. 14-day free trial. Self-service via Stripe customer portal. - [FAQ](https://docs.stackspend.app/faq): Common questions about providers, data, pricing, security, and features.