@tuned-tensor/cli 0.13.0 → 0.13.1-beta.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (4) hide show
  1. package/README.md +40 -24
  2. package/dist/index.js +481 -168
  3. package/dist/index.js.map +1 -1
  4. package/package.json +1 -1
package/README.md CHANGED
@@ -79,33 +79,41 @@ The interactive agent harness and conversation state run on your laptop.
79
79
  Inference runs through the provider/model you select, which may be a
80
80
  local endpoint or a remote provider.
81
81
 
82
- `tt` reuses provider authentication and custom model definitions from
83
- `~/.pi/agent/auth.json` and `~/.pi/agent/models.json`. Authenticate a provider
84
- through its normal login flow; `tt` deliberately has no provider-secret flags.
85
- Then inspect and select a provider/model:
82
+ Run `tt` to open one terminal for conversation and commands. Workflow commands
83
+ such as `runs list` and `doctor` work immediately. Chat waits until you choose
84
+ a provider and model from inside the shell:
86
85
 
87
- ```bash
88
- tt agent models --all
89
- tt agent configure --provider anthropic --model claude-sonnet-4-5 --thinking high
90
- tt agent status
91
- ```
86
+ ```text
87
+ tt v0.13.1-beta.1
88
+ agent not configured · workflow model base
89
+ ctrl+c stop/clear · ctrl+d exit · /help commands · tab complete
92
90
 
93
- `--thinking` accepts `off`, `minimal`, `low`, `medium`, `high`, `xhigh`, or
94
- `max`.
95
- Selection metadata can be overridden per process with
96
- `TUNED_TENSOR_AGENT_PROVIDER`, `TUNED_TENSOR_AGENT_MODEL`, and
97
- `TUNED_TENSOR_AGENT_THINKING`. These values are not credentials.
91
+ Use /model to choose a provider and model. Workflow commands work now.
92
+ › /model
93
+ › /login openai
94
+ › /model openai
95
+ › /model openai/gpt-5.6-sol
96
+ ```
98
97
 
99
- After configuration, run `tt` to open one terminal for conversation and
100
- commands:
98
+ Use `/login` to save a provider API key. The shell asks which provider, then
99
+ prompts for the key with hidden input and stores it under TT's local agent
100
+ config. `/login <provider>` skips the provider prompt.
101
+ `tt` does not accept provider secrets as flags and does not read
102
+ `~/.pi/agent/`. `/model` and `/login` show OpenAI and OpenRouter.
103
+ Other catalog providers remain available if you type their id.
104
+ Unauthenticated entries are marked `auth required`. Local endpoints such as
105
+ Ollama still need a placeholder `apiKey` in `models.json`.
101
106
 
102
107
  Before the shell opens, `tt` performs a short, non-blocking npm version check.
103
108
  If a newer stable release is available it recommends
104
109
  `npm install -g @tuned-tensor/cli@latest`; offline or unavailable registry
105
110
  checks are ignored.
106
111
 
112
+ After a model is selected, the banner shows it and ordinary sentences go to
113
+ the agent:
114
+
107
115
  ```text
108
- tt v0.13.0
116
+ tt v0.13.1-beta.1
109
117
  agent anthropic/claude-sonnet-4-5 · workflow model base
110
118
  ctrl+c stop/clear · ctrl+d exit · /help commands · tab complete
111
119
 
@@ -138,16 +146,24 @@ is bound to the current workspace so changing directories before approval fails
138
146
  safely instead of writing somewhere else. Ambiguous write failures remain
139
147
  sealed as `outcome_unknown` for manual inspection. `/reject` never mutates.
140
148
 
141
- Useful shell controls include `/help`, `/status`, `/context`, `/model`, `/cd`,
142
- `/clear`, and `/exit`.
149
+ Useful shell controls include `/help`, `/status`, `/context`, `/model`,
150
+ `/login`, `/cd`, `/clear`, and `/exit`.
143
151
  The shell banner and `/context` surface the assistant's configured provider and
144
152
  model (`agent provider/model`) separately from the workflow model, so the
145
153
  model answering your prompts is always visible. `/model` shows and changes the
146
- laptop-local TT agent model: run it with no arguments for a short list of
147
- suggestions, search with `/model <query>` (only the closest matches are
148
- shown), or switch with `/model <provider>/<model>` (for example
149
- `/model anthropic/claude-sonnet-4-5`). The equivalent non-interactive commands
150
- are `tt agent models`, `tt agent configure`, and `tt agent status`.
154
+ laptop-local TT agent model: run it with no arguments to list providers and the
155
+ two suggested models (GPT-5.6 Sol and DeepSeek V4 Flash), `/model <provider>`
156
+ to list that provider's models, search
157
+ with `/model <query>` (only the closest matches are shown), or switch with
158
+ `/model <provider>/<model>` (for example `/model anthropic/claude-sonnet-4-5`).
159
+ `/login` asks which provider to authenticate, then prompts for that provider's
160
+ API key with hidden input. `/login <provider>` skips the provider prompt.
161
+ The equivalent non-interactive commands are `tt agent models`,
162
+ `tt agent configure`, and `tt agent status`.
163
+ `tt agent configure --thinking` accepts `off`, `minimal`, `low`, `medium`,
164
+ `high`, `xhigh`, or `max`. Selection metadata can be overridden per process with
165
+ `TUNED_TENSOR_AGENT_PROVIDER`, `TUNED_TENSOR_AGENT_MODEL`, and
166
+ `TUNED_TENSOR_AGENT_THINKING`. These values are not credentials.
151
167
  The shell keeps normal terminal scrollback and command history only for the
152
168
  current process.
153
169