Pension data from Company A as structured JSON, with value and currency.

curl https://api.anessa.co/v1/accounts \
  -H "Authorization: Bearer $TOKEN"

Editor note (hidden on the live site): paste code in the Code field on the right and set the card header in Code Title / Code Meta (leave both empty to hide it). Lines starting with // are comments; a quoted line that isn't a key (a chat question or answer) shows as text. Colours and line numbers only appear on the published page.

No more guesswork

Without real data, AI answers about money are generic at best and wrong at worst. Your customers’ financial context lives with other providers. We bring it to your AI.

Financial context for AI

Chat where a user asks what assets they have, and the answer via Anessa MCP shows 300 000 EUR across six providers

Access data across providers

Give your AI the investments, pensions and savings customers hold with banks, pension providers and platforms.

Calculated total annual cost split into fund fees and platform fee, compared with the peer median

Deterministic, not generated

Fees, composition, concentration and benchmarks are calculated from the data, never generated by a model, so your AI works with exact figures.

Example MCP tool calls with MCP and API options

Connect through MCP or API

Plug into assistants and agents through MCP, or build it into your own product through our API.

What you can build

Start with a question your customers already ask.

AI assistant answering "What assets do I have, and where?" with pensions, funds and savings across six providers in EUR.

curl https://api.anessa.co/v1/accounts \
  -H "Authorization: Bearer $TOKEN"

Editor note (hidden on the live site): paste code in the Code field on the right and set the card header in Code Title / Code Meta (leave both empty to hide it). Lines starting with // are comments; a quoted line that isn't a key (a chat question or answer) shows as text. Colours and line numbers only appear on the published page.

See it in your personal assistant

Connect the Anessa MCP to an assistant that supports MCP and ask a real question about your finances.

Test yourself

FAQ

What is MCP?

A standard that lets an AI assistant or agent query an external data source conversationally, without the builder constructing that integration layer. Anessa runs one for financial data.

Which assistants does it work with?

Any assistant or agent that supports MCP including Claude, ChatGPT, Gemini, Mistral, Microsoft Copilot, Grok and Perplexity.

Do we have to use your consent interface?

No, but we recommend it. Anessa Connect, our consent interface, handles provider selection, authentication, consent and error handling across every provider you support. That part is slow and complex to build well, and it keeps changing as providers update their login flows. If you’d rather build your own, the same data is available directly through our Financial data API.

Does Anessa host models?

No. Your large language model runs where it does today, hosted by you or your AI provider. Anessa provides the building blocks: consented financial data and calculated insights such as fees, portfolio analysis and benchmarks. Your model works from exact figures instead of estimates.

Is it secure and consented?

Yes. Nothing is collected until the customer has given their consent and authorised Anessa’s assisted access to their financial service provider. Anessa uses the customer’s credentials only to carry out the data collection and discards them immediately afterwards. They are never shared and never reach your systems. All data is encrypted in transit and at rest and stored in the EU, and Anessa is ISO 27001 certified.