Right now, every AI chat on your team starts from zero — and every repo carries its own ageing rule file. Give them all one encrypted source of truth instead: engineers stay in flow, new hires onboard in minutes, and when an auditor asks who read what through context-mcp, the answer is one log away.
brain · context · tools
proof
We run our own medical-software company on taip: a regulated codebase running daily under MDR and IEC 62304, since 2026. Your team gets the stack we trust with our own compliance.
the founder, Taip Systems OÜ
the product
Today: Cursor invents an answer from whatever context it finds — docs that may be out of date, rules that vary by repo. With taip: only the documentation their role is permitted to read, decrypted in memory. Everything else stays encrypted and out of the prompt.
Today: evidence is assembled by hand — searching logs, chasing repo owners, hoping nothing was missed. With taip: context-mcp records who read what under which role, and a subject's data can be crypto-shredded on request. Minutes, not days.
Today: access tokens are revoked one by one — if anyone remembers which repos they could read. With taip: revoking the key ends access to TAIP-managed content. No plaintext IP is left behind in taip's own storage. Data in connected AI providers' chat history is outside taip's control.
why taip
Under Regulation (EU) 2024/1689, Annex III high-risk provisions apply from 2027-12-02 and Annex I product-embedded provisions from 2028-08-02. taip supports selected technical and organisational controls relevant to these requirements — encryption at rest, role-based access, and context-mcp access logs from the start, not configured per repository.
Company knowledge is encrypted at rest and decrypted only in memory, after role-based authorization checks plaintext frontmatter.
context-mcp records every read and write with actor, role, item, and timestamp. Access records for the assistant sessions that matter most.
Each company gets its own tenant, hosted in the EU and operated under EU law.
trust & safety
Content is encrypted at rest; role filtering runs before decryption. Revoking the key ends access to TAIP-managed content; destroying it triggers crypto-shred for GDPR erasure.
Authorization runs on plaintext frontmatter before any decryption, and fails closed. A role reads only what it's permitted; everything else stays ciphertext.
Hosted in the EU and operated under EU law. Designed for GDPR and EU AI Act requirements, with context-mcp access logs.
Grant a new teammate the right rules and context in minutes; revoke a leaver by revoking their key, ending access to TAIP-managed content.
alpha access
We onboard companies one pilot at a time and work with each directly. No spam. Unsubscribe any time.