Powering privacy infrastructure for building payments into apps, bots and AI agents in Python. Built on @zcash, completely Open-source
Joined October 2026
- Tweets2
- Following3
- Followers23
- Likes39
Zecpy retweeted
Software is how money moves today, from SaaS subscriptions and API micro-billing to automated payroll and autonomous AI agents buying compute on demand.
Every standard payment leaves a metadata footprint detailing who paid, how much, and how often.
Python handles a massive portion of this financial infrastructure, but developers have never had a simple, native way to process private payments.
ZecPy solves this.
By running pip install zecpy, any Python backend, script, or AI agent can send and receive shielded Zcash in just a few lines of code.
It handles key tasks out of the box: issuing unique private addresses for every invoice, managing bulk payouts to dozens of recipients in a single transaction, supporting cold-storage offline signing, and setting hard budget caps for autonomous agents. Under the hood, it relies directly on Zcash's official Rust cryptography (librustzcash).
Privacy on-chain does not automatically guarantee end-to-end privacy. Standard wallets query light servers to sync with the blockchain, and those network requests routinely expose identifying metadata including wallet addresses, transaction histories, birthday heights, and sync states. The blockchain itself stays shielded, but the light server learns your identity.
ZecPy is built from the ground up to prevent metadata leaks.
To verify this, we built a telemetry proxy that monitors and logs every outbound request during routine wallet tasks.
In side-by-side testing across five standard steps, two established Zcash wallets made 7 and 71 identifying requests respectively. ZecPy made zero.
In a full e-commerce simulation over Tor on testnet, ZecPy generated 108 requests with no identifying leaks, outside of one optional memo fetch that masked the transaction inside a batch of decoy payments. Our benchmark testing also uncovered three upstream bugs in core Zcash repositories, which we filed with fixes (librustzcash #3198, #3199; zcash-devtool #267).
This is fully functional software backed by multiple unit tests on a local devnet, real proofs, real blocks, and published testnet transactions. It is currently an alpha release and has not yet undergone a third-party security audit.
- Try the live request telemetry demo: zecpy-live.pages.dev
- Read the documentation and AI-agent specs: zecpy.pages.dev
Software is how money moves today, from SaaS subscriptions and API micro-billing to automated payroll and autonomous AI agents buying compute on demand.
Every standard payment leaves a metadata footprint detailing who paid, how much, and how often.
Python handles a massive portion of this financial infrastructure, but developers have never had a simple, native way to process private payments.
ZecPy solves this.
By running pip install zecpy, any Python backend, script, or AI agent can send and receive shielded Zcash in just a few lines of code.
It handles key tasks out of the box: issuing unique private addresses for every invoice, managing bulk payouts to dozens of recipients in a single transaction, supporting cold-storage offline signing, and setting hard budget caps for autonomous agents. Under the hood, it relies directly on Zcash's official Rust cryptography (librustzcash).
Privacy on-chain does not automatically guarantee end-to-end privacy. Standard wallets query light servers to sync with the blockchain, and those network requests routinely expose identifying metadata including wallet addresses, transaction histories, birthday heights, and sync states. The blockchain itself stays shielded, but the light server learns your identity.
ZecPy is built from the ground up to prevent metadata leaks.
To verify this, we built a telemetry proxy that monitors and logs every outbound request during routine wallet tasks.
In side-by-side testing across five standard steps, two established Zcash wallets made 7 and 71 identifying requests respectively. ZecPy made zero.
In a full e-commerce simulation over Tor on testnet, ZecPy generated 108 requests with no identifying leaks, outside of one optional memo fetch that masked the transaction inside a batch of decoy payments. Our benchmark testing also uncovered three upstream bugs in core Zcash repositories, which we filed with fixes (librustzcash #3198, #3199; zcash-devtool #267).
This is fully functional software backed by multiple unit tests on a local devnet, real proofs, real blocks, and published testnet transactions. It is currently an alpha release and has not yet undergone a third-party security audit.
- Try the live request telemetry demo: zecpy-live.pages.dev
- Read the documentation and AI-agent specs: zecpy.pages.dev