ChatGPT is great at writing, calculating and advising \u2014 but it doesn’t know your accounts. Ask it about unpaid invoices and it will answer with generalities, because it cannot see into POHODA. The good news: connecting ChatGPT to POHODA is possible today, even without a developer. In this guide we walk through the three possible routes, what you need for each and what to watch out for.<\/p>\n
The simplest way to start: export a report from POHODA and paste it into the chat. It works immediately and costs nothing \u2014 which is why most companies start there. But there are three catches: it is laborious<\/strong> (every question = a new export), the data is stale<\/strong> the moment you paste it, and above all you lose control<\/strong> over what exactly leaves for the chat. Fine for a one-off analysis; not for daily work.<\/p>\n Community MCP servers for POHODA exist on GitHub, communicating over the XML interface. They are free and fully under your control \u2014 which is also the trade-off: you need a running mServer<\/a> (a dedicated POHODA instance), a machine for the server to run on, and someone to set it all up and maintain it through every POHODA update. If you have a capable IT person in-house and time to experiment, it is a legitimate way to get a feel for MCP.<\/p>\n The third route hands the worries over to a provider: the MCP server runs in the cloud, connects securely to your POHODA database (IP whitelisting or your VPN), and all you do is link ChatGPT. That is how our POHODA MCP<\/a> works \u2014 reads go straight from the database, so you don’t need mServer, and the connection takes three steps:<\/p>\n Whichever route you choose apart from manual copying, insist on answers to four questions: Does the connection run read-only by default<\/strong>? Can you restrict the areas<\/strong> the AI can see (payroll never, for example)? Is there an audit log<\/strong> of every query? And where does the data live<\/strong> \u2014 does it stay in the EU, and is it kept out of model training? A serious solution answers yes four times.<\/p>\n In the GPT catalogue you will also find \u201cPOHODA guide\u201d-style assistants. They can help you operate the program \u2014 explain where to find a report or how to post a document. But they cannot see your data<\/strong>: they won’t tell you who owes what, nor prepare your VAT documentation. For that you need a real connection via MCP \u2014 which is what this article is about.<\/p>\n<\/div>\n Yes \u2014 MCP is an open standard, so one MCP server works with every assistant that supports it. Claude connects in much the same way (connectors in the settings); Microsoft Copilot support is on its way.<\/p>\n<\/details>\n Not for reading data \u2014 the hosted service reads directly from the database of the SQL\/E1 editions. mServer comes into play for writes (issuing an invoice) and for the open-source route, which communicates through it. Details in our mServer article<\/a>.<\/p>\n<\/details>\n ChatGPT only receives the data the MCP server returns for a specific query \u2014 never the whole database. On the business plans (Team\/Enterprise), conversations are not used to train models. The MCP server itself should run in the EU and store no data; ask your provider about that.<\/p>\n<\/details>\n You pay the ChatGPT subscription to OpenAI. Our hosted MCP server for POHODA starts at 2 490 K\u010d per month per company \u2014 see the current pricing and what it includes on the POHODA MCP<\/a> page.<\/p>\n<\/details>\n","protected":false},"excerpt":{"rendered":" Three ways to connect ChatGPT to POHODA \u2014 from manual copy-paste to an MCP server. What you need, what it costs and what to watch out for.<\/p>\n","protected":false},"author":0,"featured_media":67149,"template":"","class_list":["post-71877","erpedie","type-erpedie","status-publish","has-post-thumbnail","hentry"],"acf":[],"_links":{"self":[{"href":"https:\/\/symmy.com\/en\/wp-json\/wp\/v2\/erpedie\/71877","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/symmy.com\/en\/wp-json\/wp\/v2\/erpedie"}],"about":[{"href":"https:\/\/symmy.com\/en\/wp-json\/wp\/v2\/types\/erpedie"}],"version-history":[{"count":1,"href":"https:\/\/symmy.com\/en\/wp-json\/wp\/v2\/erpedie\/71877\/revisions"}],"predecessor-version":[{"id":71950,"href":"https:\/\/symmy.com\/en\/wp-json\/wp\/v2\/erpedie\/71877\/revisions\/71950"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/symmy.com\/en\/wp-json\/wp\/v2\/media\/67149"}],"wp:attachment":[{"href":"https:\/\/symmy.com\/en\/wp-json\/wp\/v2\/media?parent=71877"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}Route 2: an open-source MCP server (for the technically minded)<\/h2>\n
Route 3: a hosted service (no installation, live in days)<\/h2>\n
\n
https:\/\/mcp.symmy.ai\/pohoda<\/code>) and sign in \u2014 it uses OAuth, no shared keys and no POHODA passwords.<\/li>\nWhat you can do with the connection<\/h2>\n
\n
Security: what to ask before letting AI near your data<\/h2>\n
Don’t confuse the two: a \u201cGPT advisor\u201d is not a data connection<\/h2>\n
Frequently asked questions<\/h2>\n
Does the same work with Claude or Copilot?<\/summary>\n
Do I need mServer?<\/summary>\n
Will OpenAI see my accounts?<\/summary>\n
What does it all cost?<\/summary>\n