# Ithaca ## Docs - [MCP Connection](https://docs.doomberg.me/agent/connection.md): How agents connect to Ithaca over MCP — stdio vs streamable-HTTP transports, bearer-token authentication, the initialize handshake, tool discovery, resource URIs, and the session lifecycle. - [Starter Prompt](https://docs.doomberg.me/agent/starter-prompt.md): The starter_research MCP prompt — its parameters, the full template text it returns, and ready-to-paste example prompts for your agent. - [Agent Workflow](https://docs.doomberg.me/agent/workflow.md): The canonical research workflow the agent follows — session_context, data tools, propose_strategy, backtest_run, subscribe_run, report — with server instructions and a full NVDA example session. - [Approval endpoints](https://docs.doomberg.me/api/approvals.md): List and decide promotion approvals — agents request, humans decide, every step is audited. - [REST API overview](https://docs.doomberg.me/api/overview.md): Base URL, authentication, tenant isolation, response format, and the full endpoint catalog for the Ithaca REST API. - [Provider endpoints](https://docs.doomberg.me/api/providers.md): List and manage data provider connections and dataset snapshots — credentials are never exposed. - [Run endpoints](https://docs.doomberg.me/api/runs.md): Create, inspect, cancel, and promote backtest and paper-trading runs — the core execution surface of Ithaca. - [Session endpoints](https://docs.doomberg.me/api/sessions.md): List, create, and seed demo research sessions — the top-level container for traced agent activity. - [Strategy endpoints](https://docs.doomberg.me/api/strategies.md): List and propose immutable, versioned strategy specs — the declarative inputs to backtests. - [Trigger endpoints](https://docs.doomberg.me/api/triggers.md): Create, manage, and dry-run quote-based and time-based triggers that drive automated agent sessions. - [Introduction](https://docs.doomberg.me/index.md): Ithaca is a quant research, backtest, and paper-trading observability platform where AI agents call tools over MCP to research markets and run strategies on trusted infrastructure. - [Quickstart](https://docs.doomberg.me/quickstart.md): Connect your agent to Ithaca in one command — approval, key minting, shell sourcing, and agent registration. - [Computed Analytics Skills](https://docs.doomberg.me/skills/computed.md): Reference for correlation.matrix, factor.exposures, dcf.value, portfolio.optimize, montecarlo.simulate, scenario.project, risk.stress, tearsheet, walkforward.run, technical.indicators, seasonality.monthly, and seasonality.heatmap — the computed analytics skills exposed to AI agents. - [Congress & Insider Skills](https://docs.doomberg.me/skills/congress-insider.md): Reference for congress.trades, congress.late_filings, congress.screener, insider.activity, and politician.portfolio — the archive-backed political-data skills exposed to AI agents. - [Financials & Earnings Skills](https://docs.doomberg.me/skills/financials.md): Reference for financials.income, financials.balance, financials.cashflow, earnings.transcript, earnings.calendar, dividends, and stock.splits — the archive-backed financials and earnings skills exposed to AI agents. - [Fundamentals Skills](https://docs.doomberg.me/skills/fundamentals.md): Reference for fundamentals.get, fundamentals.compare, fundamentals.pit, and analyst.ratings — the archive-backed fundamentals skills exposed to AI agents. - [Institutional & ETF Skills](https://docs.doomberg.me/skills/institutional.md): Reference for institutional.holdings, institutional.flows, etf.holdings, and etf.exposure — the archive-backed institutional and ETF skills exposed to AI agents. - [Market Data Skills](https://docs.doomberg.me/skills/market-data.md): Reference for market.news, market.screener, market.movers, market.quote, and market.tide — the archive-backed market data skills exposed to AI agents. - [Options Skills](https://docs.doomberg.me/skills/options.md): Reference for options.greeks and options.implied_vol — pure computed Black-Scholes options skills exposed to AI agents. - [Skills Overview](https://docs.doomberg.me/skills/overview.md): Skills are archive-backed data tools and computed analytics exposed to AI agents over MCP. Learn what they are, how they differ from direct tools, and how to discover them. - [Sentiment & Macro Skills](https://docs.doomberg.me/skills/sentiment-macro.md): Reference for sentiment.get, macro.dashboard, universe.list, economic.calendar, fda.calendar, commodity.prices, forex.rates, crypto.prices, and vix.term_structure — the archive-backed sentiment and macro skills exposed to AI agents. - [Short Interest Skills](https://docs.doomberg.me/skills/short-interest.md): Reference for short.interest, short.volume, short.ftd, and short.screener — the archive-backed short interest skills exposed to AI agents. - [Volatility Skills](https://docs.doomberg.me/skills/volatility.md): Reference for vol.realized, vol.term_structure, vol.iv_rank, vol.vrp, vol.anomaly_score, and vol.character — computed-from-archive-closes volatility skills exposed to AI agents. - [Approval workflow](https://docs.doomberg.me/strategy/approvals.md): How a backtested strategy is promoted to paper trading — agents request, humans decide, and every step is audited. - [Artifacts](https://docs.doomberg.me/strategy/artifacts.md): The 8-file backtest artifact bundle — schemas, content hashing, storage backends, and the REST API for retrieval. - [Backtest system](https://docs.doomberg.me/strategy/backtest.md): How backtest_run turns a StrategySpec into a deterministic equity curve, metrics bundle, and risk attestation. - [StrategySpec](https://docs.doomberg.me/strategy/spec.md): The full declarative strategy schema — a closed, versioned contract that agents propose and the backtest engine executes. - [Approval Tools](https://docs.doomberg.me/tools/approvals.md): list_approvals and decide_approval are the human's promotion workflow. The agent can request paper-trading promotion, but only a human can approve it. - [Artifact Tools](https://docs.doomberg.me/tools/artifacts.md): get_artifact fetches artifact metadata and a download URI. Every backtest produces an 8-file bundle with SHA-256 content hashing for reproducibility. - [Backtest Tools](https://docs.doomberg.me/tools/backtest.md): backtest_run runs a deterministic backtest over a StrategySpec and date range, returning an equity curve, full stats, final weights, and a risk attestation. Same inputs always produce the same hash. - [Data Tools](https://docs.doomberg.me/tools/data.md): market_get_ohlcv fetches price data with a provenance envelope; search_skills and describe_skill power progressive discovery of the 50+ skill-backed data sources. - [Tools Overview](https://docs.doomberg.me/tools/overview.md): The full MCP tool catalog at a glance — 13 direct tools plus 50+ skill-backed tools, grouped by category. - [Run Management Tools](https://docs.doomberg.me/tools/runs.md): run_skill, get_run, subscribe_run, and cancel_run manage the durable run lifecycle — draft, queued, running, succeeded, failed, cancelled. - [Session Tools](https://docs.doomberg.me/tools/session.md): session_context opens a traced research run, attaches the user's question, and scopes every subsequent tool call. It is always called first. - [Strategy Tools](https://docs.doomberg.me/tools/strategy.md): propose_strategy validates a declarative StrategySpec against the frozen contract, stamps tenant_id, and stores an immutable versioned row. No agent-authored code, ever. ## Optional - [Web App](https://app.doomberg.me) - [GitHub](https://github.com/joshuajerin/doomberg)