Local Studio workflow

tezz ai production --root examples --root lib --root TezzAi --root tools --data build/ai_prod/samples.tnxb --model build/ai_prod/model.taim --codebook build/ai_prod/codebook.tnxb
tezz ai serve --host 127.0.0.1 --port 8099 --model build/ai_prod/model.taim --codebook build/ai_prod/codebook.tnxb --v2

Open UI:
http://127.0.0.1:8099/

UI features:
- multi-session chat ids
- code-only response mode
- session export JSON
- template controls (auto/cli/api/lib/service)

Core CLI commands

tezz ai doctor
tezz ai learn-corpus --root examples --root lib --root TezzAi --root tools --data build/ai_prod/samples.tnxb --model build/ai_prod/model.taim --codebook build/ai_prod/codebook.tnxb --no-train
tezz ai train --data build/ai_prod/samples.tnxb --model build/ai_prod/model.taim --v2
tezz ai code "build api websocket route with auth middleware" --model build/ai_prod/model.taim --codebook build/ai_prod/codebook.tnxb --out build/ai_prod/gen_api.tn --json-out build/ai_prod/gen_api.json --pick-debug --v2

Serve endpoint

tezz ai serve --host 127.0.0.1 --port 8099 --model build/ai_prod/model.taim --codebook build/ai_prod/codebook.tnxb --v2

GET  /health
GET  /ai/templates
GET  /ai/code?prompt=...
POST /ai/code (form/json: prompt, template, session_id)
POST /ai/chat (json: prompt, template, session_id, max_code)

Production guarantees

- Model format: .taim (secure payload lane)
- Dataset format: .tnxb
- Source-pick controls: --pick-debug and explicit blocker policies
- Default ranking bias prefers examples/lib over tests for clean generation
- Session-safe responses for frontend/web clients

Quick links

CLI install:
/download/

Getting started:
/docs/getting-started

Framework stack:
/frameworks/tezzapi
/frameworks/tezzserve
/frameworks/tezzdb