@tduyng/pi-commandcode 0.0.2 → 0.0.3

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 (2) hide show
  1. package/README.md +34 -16
  2. package/package.json +2 -2
package/README.md CHANGED
@@ -1,10 +1,16 @@
1
1
  # pi-commandcode
2
2
 
3
- Command Code models in Pi through the documented Provider API. Unofficial: this extension is not made or endorsed by Command Code.
3
+ Command Code models in pi through the documented Provider API, with pi's own prompt, tools, and approvals. Unofficial: not made or endorsed by Command Code.
4
+
5
+ ## What it costs
6
+
7
+ No tools, no prompt block, no schema. It is a model provider, so it adds nothing to the context and nothing to the cold start.
8
+
9
+ Streaming, tool calls, and message conversion are pi's own adapters, so the transport adds no round trip of its own.
4
10
 
5
11
  ## Requirements
6
12
 
7
- - A Command Code API key from a plan with API access: GOAT, Pro, Max, Team or Provider. The Go and $1 plans have no API access.
13
+ - An API key from a plan with API access: GOAT, Pro, Max, Team, or Provider. The Go and $1 plans have no API access.
8
14
  - Pi 1.4 or newer.
9
15
 
10
16
  ## Use
@@ -14,7 +20,7 @@ export COMMAND_CODE_API_KEY="user_..."
14
20
  pi --provider commandcode --model claude-opus-5
15
21
  ```
16
22
 
17
- Inside Pi, `/model` lists every `commandcode` model. The list is fetched from `https://api.commandcode.ai/provider/v1/models` at startup, and again each time the model selector opens.
23
+ Inside pi, `/model` lists every `commandcode` model. The list is fetched from `https://api.commandcode.ai/provider/v1/models` at startup and again each time the selector opens.
18
24
 
19
25
  ## How it works
20
26
 
@@ -23,27 +29,39 @@ Inside Pi, `/model` lists every `commandcode` model. The list is fetched from `h
23
29
  | `claude-*` | Anthropic Messages, `/provider/v1/messages` | `https://api.commandcode.ai/provider` |
24
30
  | everything else | OpenAI Chat Completions, `/provider/v1/chat/completions` | `https://api.commandcode.ai/provider/v1` |
25
31
 
26
- Streaming, tool calls and message conversion are Pi's own adapters. The extension sends the key as a Bearer token for Chat Completions and as `x-api-key` for Messages, as the Provider API documents.
32
+ The key goes as a Bearer token for Chat Completions and as `x-api-key` for Messages, as the Provider API documents. `typesafe/*` decision models are left out because they run on a separate endpoint and never stream.
27
33
 
28
- `typesafe/*` decision models are left out because they run on a separate endpoint and never stream.
34
+ ## Limits
29
35
 
30
- ## Configuration
36
+ - The catalog carries no image or price metadata, so every model is text-only and shows zero cost in pi.
31
37
 
32
- | Variable | Meaning |
33
- | --- | --- |
34
- | `COMMAND_CODE_API_KEY` | Provider API key. Required for requests. |
35
- | `CMD_ZDR=1` | Sends `x-cmd-zdr: 1`, so requests go only to zero-data-retention upstreams. A model without one fails with HTTP 422. |
38
+ - Thinking is on for every model. Pi levels map to `reasoning_effort`, and `minimal` sends `low`. `off` is offered only for the DeepSeek ids that accept it, because a model that cannot stop thinking rejects `off` with HTTP 400.
36
39
 
37
- ## Limits
40
+ - Claude 4.6 and newer use adaptive thinking, and the 4.5 models take a thinking budget.
41
+
42
+ - Every model gets a 32768 output limit, the smallest in the Command Code CLI catalog, because per-model limits are not in the Provider API.
38
43
 
39
- - The catalog has no image or price metadata, so every model is text-only and shows zero cost in Pi.
40
- - Thinking is on for every model. Pi levels map to the gateway's `reasoning_effort` values, `minimal` sends `low`. `off` is offered only for the DeepSeek ids that accept it, because a model that cannot stop thinking rejects `off` with HTTP 400 and leaves the level out everywhere else.
41
- - Claude 4.6 and newer use adaptive thinking, the 4.5 models take a thinking budget.
42
- - Every model gets a 32768 output limit, the smallest limit in the Command Code CLI catalog. Per-model limits are not in the Provider API.
43
44
  - Chat Completions is used for every non-Claude model, as the docs say. A model that rejects that wire returns its own error.
44
- - Messages requests carry Pi's `?beta=true` query string, added by the Anthropic SDK.
45
+
46
+ - Messages requests carry pi's `?beta=true` query string, added by the Anthropic SDK.
47
+
45
48
  - If no models appear, check the network and the key, then open `/model` again.
46
49
 
50
+ ## Configure
51
+
52
+ | Variable | Meaning |
53
+ | --- | --- |
54
+ | `COMMAND_CODE_API_KEY` | Provider API key, required for requests |
55
+ | `CMD_ZDR=1` | Sends `x-cmd-zdr: 1`, so requests go only to zero-data-retention upstreams. A model without one fails with HTTP 422 |
56
+
57
+ ## Install
58
+
59
+ ```bash
60
+ pi install npm:@tduyng/pi-commandcode
61
+ ```
62
+
63
+ Local checkout: `pi install /path/to/dotfiles/pi/agent/extensions/pi-commandcode`
64
+
47
65
  ## Tests
48
66
 
49
67
  ```sh
package/package.json CHANGED
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "name": "@tduyng/pi-commandcode",
3
- "version": "0.0.2",
4
- "description": "Use Command Code models in Pi through the documented Provider API, with Pi's own prompt and tools.",
3
+ "version": "0.0.3",
4
+ "description": "Command Code models in Pi through the documented Provider API, with Pi's own prompt and tools.",
5
5
  "keywords": [
6
6
  "pi",
7
7
  "pi-package",