Use nordcap with your AI

749 Nordic small-caps — fundamentals, news, insider trades — wired straight into the assistant you already use. Pick your platform; every guide below is copy-paste-ready. You'll need a free API key for everything except the Claude MCP connector.

Claude MCP + Skills

Connector (no key needed): Claude.ai / Desktop → Settings → Connectors → Add custom connector → paste https://mcp.nordcap.dev/mcp (leave advanced fields empty). Claude Code:

claude mcp add --transport http nordcap https://mcp.nordcap.dev/mcp

Agent Skills (Claude Code): the data skill teaches the agent the full API, webhooks and coverage rules; The Council convenes a six-seat AI investment committee that debates any ticker over live data.

mkdir -p ~/.claude/skills/nordcap ~/.claude/skills/nordcap-council
curl -sL https://nordcap.dev/skill/SKILL.md -o ~/.claude/skills/nordcap/SKILL.md
curl -sL https://nordcap.dev/skill/council/SKILL.md -o ~/.claude/skills/nordcap-council/SKILL.md

Then: "Convene the council on CURAS." Full details in the docs.

ChatGPT Custom GPT + Actions

ChatGPT calls nordcap through a Custom GPT with Actions (requires a ChatGPT plan with GPT building):

1. ChatGPT → Explore GPTs → Create → Configure tab
2. Under Actions → Create new action → Import from URL → paste https://api.nordcap.dev/openapi.json
3. Authentication → API Key → Auth type Custom → header name X-API-KEY → paste your key from the dashboard
4. Paste the instructions below into the GPT's Instructions field and save

Instructions — nordcap analyst GPT

You are a Nordic small-cap research analyst with live access to the nordcap API (749 companies on First North, Spotlight, NGM, Euronext Growth — the companies US data vendors can't see).

Rules:
- For ANY question about a Nordic small-cap, call the API first. Never answer from memory.
- Endpoints: /tickers (universe), /company/facts, /financials/income-statements (annual, DK up to 13y + Norwegian accounts), /news, /insider-trades (MAR art. 19), /filings.
- Swedish/Finnish companies have facts + news + insider trades but NO financial statements yet — say so plainly and analyze with what exists.
- nordcap has NO share prices. Never invent one; ask the user or analyze without.
- Always state the reporting currency (DKK/NOK/SEK/EUR) and the period_end of any figures, and how old they are.
- Insider trades are strong small-cap signals — check them on every deep dive.
- Everything is information, never a buy/sell recommendation. Say so when the user asks for one.
- Large caps (Novo Nordisk etc.) are NOT covered — they live on main markets; suggest mainstream data sources for those.

Bonus: The Council as a GPT

Create a second GPT with the same Action, and paste the full Council playbook as its instructions — it fits:

curl -sL https://nordcap.dev/skill/council/SKILL.md

(Open the URL, copy the markdown body into the Instructions field. The six personas, the cross-examination round and the verdict card all work in ChatGPT.)

Perplexity Limited

Honest answer: Perplexity can't send custom auth headers, so it cannot call the nordcap API directly. What works today:

• Spaces: create a Space ("Nordic small-caps") and paste the instructions below — Perplexity will then use nordcap's public pages for coverage questions and tell you exactly what data to fetch yourself.
• Paste-and-analyze: run a curl from the docs, paste the JSON into the thread — the instructions teach it how to read our shapes.

You help me research Nordic small-cap stocks (First North, Spotlight, NGM, Euronext Growth). My data source is nordcap.dev — coverage and field reference at nordcap.dev/docs.

When I ask about a Nordic small-cap: 1) check nordcap.dev/docs for what data exists (Danish/Norwegian financials, pan-Nordic news and MAR insider trades; no share prices), 2) give me the exact curl command to fetch it (base https://api.nordcap.dev, header X-API-KEY), 3) when I paste JSON back, analyze it: state currency and period_end, compute cash runway for loss-makers, flag stale accounts and insider activity. Never present analysis as investment advice.

Any agent framework REST + markdown

Building with LangChain, CrewAI, the OpenAI Agents SDK, or your own stack? Everything is plain HTTP and markdown:

• OpenAPI spec: https://api.nordcap.dev/openapi.json — generate a typed client or feed it to a tool-calling agent
• Skill files (system-prompt-ready markdown): data skill · The Council
• MCP (streamable HTTP): https://mcp.nordcap.dev/mcp — works with any MCP client
• Webhooks to wake your agent on new events: docs (Quant tier)

Missing your platform? Tell us — integrations ship fast here.