Roadmap

This is a direction, not a set of dated commitments. It is priorities pulled by real deployments, and it exists so you can see where TokenVeil is heading before you build on it. Items here are in development or planned, not shipped. What is shipped is documented in the rest of this site.

Transparent gateway for coding agents

A drop-in gateway for Claude Code and Codex: you point the tool at TokenVeil as its model endpoint, and the full model-bound context is anonymized on the wire before it reaches the provider and restored in the reply, with no UI and no manual step. Because a gateway on the request sees everything the agent sends to the model, this covers prompts, files the agent reads, git diffs, terminal output, tool-call and MCP results, and context the agent generates autonomously. Combined with locked egress, it is the fully transparent, fail-closed form of what the API delivers today. See deployment security.

Signed Privacy Receipts

A per-transformation receipt documenting the privacy profile, the engine and version, the transformation applied and the authorized downstream uses, signed so it can be verified independently. It turns each anonymization into an auditable compliance artifact.

Configurable pseudonym scope

A configurable scope key so the same real value can receive a stable token within one workspace or case and a different token in another context, beyond the current per-session scope. See how reversibility works.

Language and detection packs

Additional language packs (a national NER model plus that country’s identifier recognizers) built on demand for specific markets. See extending detection.