TallyPrime and Anaplan integration

What moves between them

The main flow runs from TallyPrime into Anaplan. On the schedule you set, ml-connector exports general ledger balances and posted vouchers from TallyPrime over the local agent for the chosen period, formats them into the file an Anaplan import action expects, uploads the chunks, and runs that import so actuals land in the planning model. Ledger and Group masters move the same direction so the Anaplan accounts and cost center lists stay aligned with the TallyPrime chart. Approved budget or forecast figures can flow back from an Anaplan export action into TallyPrime as ledger reference data. Cadence is typically each night or each period close, since neither system pushes events.

How ml-connector handles it

ml-connector stores both credential sets encrypted and reaches TallyPrime only through the customer local agent, because port 9000 is never exposed to the cloud, posting XML or JSON envelopes with the company name in SVCURRENTCOMPANY and the date window in SVFROMDATE and SVTODATE. On the Anaplan side it logs in for an AnaplanAuthToken and reuses it within the 35 minute window rather than minting one per call, then runs the model builder's named import and export actions, polling each taskId until COMPLETE and downloading export chunks. Both systems are pull only, so it polls on a schedule rather than waiting for a push. TallyPrime Ledgers and Groups map to Anaplan Lists for accounts and cost centers so each posted value lands on a member that already exists. TallyPrime has no idempotency keys, so it tracks imported voucher numbers and uses Alter to update rather than re-import, while on Anaplan it serializes actions against a model because a running task locks it, and it backs off on the 600 request per minute limit when a call returns HTTP 429. Every record carries a full audit trail and can be replayed if a downstream call fails.

A real-world example

A mid-sized distribution business with around 200 staff runs TallyPrime for accounting, inventory, and GST across three branches, and its finance team builds the annual budget and rolling forecast in Anaplan. Before the integration, an analyst exported a trial balance and voucher reports from each TallyPrime company at month end and pasted the actuals into Anaplan by hand, which delayed plan versus actual reporting by several days and introduced mapping errors when account names did not line up. With TallyPrime and Anaplan connected, ledger balances and vouchers load into the Anaplan actuals modules automatically each period, mapped to the right account and cost center lists, and approved budget figures flow back to TallyPrime. Variance reporting is ready on the first day of close instead of after a week of re-keying.

What you can do

• •Load TallyPrime general ledger balances and posted vouchers into Anaplan planning modules each period.
• •Keep Anaplan account and cost center lists aligned with TallyPrime Ledgers and Groups.
• •Return approved Anaplan budget and forecast figures to TallyPrime as ledger reference data.
• •Bridge the on-premise TallyPrime local agent and Anaplan token or OAuth2 authentication.
• •Poll both pull-only systems on your schedule, with retries and a full audit trail on every record.

Questions

Which direction does data move between TallyPrime and Anaplan?

The main flow is TallyPrime into Anaplan. General ledger balances, posted vouchers, and chart of accounts masters move from TallyPrime into Anaplan as planning actuals and list members. Approved budget or forecast figures can flow back from an Anaplan export action into TallyPrime ledgers as reference data.

Why does TallyPrime need a local agent?

TallyPrime is a desktop application whose HTTP server on port 9000 is only reachable on the same machine or LAN, with no cloud endpoint. ml-connector talks to a small local agent the customer runs next to TallyPrime, which forwards the XML or JSON envelopes to port 9000 and relays the responses. The TallyPrime application must be running with the target company open for any call to work.

How does the integration run Anaplan imports and exports?

Anaplan only runs named import and export actions a model builder has already created in the model, so ml-connector discovers those action IDs and runs them rather than writing cells directly. Each run is asynchronous, so it starts the action, polls the returned taskId until the task reports COMPLETE, and downloads file chunks for exports. Because a running task locks the model, it serializes actions against the same model and backs off when the 600 request per minute limit returns HTTP 429.