aibill 0.9.10 → 0.9.12

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 +75 -30
  2. package/package.json +3 -3
package/README.md CHANGED
@@ -1,35 +1,80 @@
1
- # aibill
1
+ # aibill — Tilden’s free AI spend CLI
2
2
 
3
- The short npm command for
4
- [ai-spend-agent](https://github.com/futurastudio/ai-spend-agent).
3
+ Inspect Claude Code and Codex activity locally, see its estimated value at
4
+ published API rates, and identify what deserves a closer look. Optional provider
5
+ reports add billing evidence; estimates and reported costs stay separate.
5
6
 
6
7
  ```bash
7
- npx aibill
8
- npx aibill init
9
- npx aibill statusline install # optional Claude Code cache-only status line
8
+ npx aibill@latest init
9
+ npx aibill@latest --full
10
10
  ```
11
11
 
12
- It runs the exact same local-first financial-accountability CLI as
13
- `npx ai-spend-agent`: connect supported coding-agent work to cost evidence,
14
- attribution, runway, and one next action. See the repository
15
- [README](https://github.com/futurastudio/ai-spend-agent#readme) for data
16
- semantics, supported providers, privacy boundaries, and limitations.
17
-
18
- The optional status line uses the same plan-aware evidence vocabulary while
19
- reading only a private aggregate cache. Bare init does not change Claude
20
- settings, and `npx aibill statusline uninstall` restores the prior user value.
21
-
22
- Connector validation (`live_verified`, `fixture_verified`, `untested`, or
23
- `failed`) and each number's financial evidence (`verified`, `estimated`,
24
- `detected_unverified`, or `missing`) are separate status axes. Run `npx aibill
25
- doctor --sources` to see both. aibill never sits in the inference path and never stores, prints, or proxies provider credentials.
26
-
27
- **Telemetry: anonymous command counts, disclosed at first run.** Your
28
- transcripts, prompts, file names, and dollar amounts stay on your machine. The
29
- CLI separately counts *which* commands run — anonymous, never your arguments,
30
- paths, content, or your email — and only after a one-time notice has been
31
- printed on an interactive run. One command turns it off: `aibill telemetry off`
32
- (or `DO_NOT_TRACK`, `CI`, or `AI_SPEND_NO_TELEMETRY`). Full scope:
33
- [`docs/TELEMETRY.md`](https://github.com/futurastudio/ai-spend-agent/blob/main/docs/TELEMETRY.md).
34
-
35
- MIT licensed.
12
+ Requires Node.js 22+. Free and MIT-licensed; no account or provider key is
13
+ needed for local analysis. `aibill` and `ai-spend-agent` run the same CLI.
14
+ Run from a project directory: `init` reviews the last 30 days of supported
15
+ machine-wide records, uses that project for local state, and seeds the private
16
+ Claude Code/Codex status cache. It does not change Claude settings or connect
17
+ Workspace. Missing evidence stays missing; `--sample` explicitly selects a demo.
18
+
19
+ Local values are API-equivalent estimates, not subscription charges or invoices.
20
+ Claude Code/Codex readers are live-verified; Gemini CLI is experimental and
21
+ fixture-verified, supported only in financial CLI/report/local MCP surfaces.
22
+ Gemini does not feed statusline, Glance, Context Health, Apply or plan/runway.
23
+ `npx aibill@latest doctor --sources` shows reader validation separately from
24
+ each number’s financial label and coverage.
25
+
26
+ ## Choose a next step
27
+
28
+ - `npx aibill@latest context`: inspect context evidence.
29
+ - `npx aibill@latest improve`: guide one reversible token experiment; a result
30
+ requires matched completed sessions and user-declared quality, not a claim
31
+ of cash savings or ROI.
32
+ - `npx aibill@latest statusline install`: optionally install the cache-only
33
+ Claude Code status line. Remove it with `statusline uninstall`.
34
+ - `npx aibill@latest connect openai` or `connect anthropic`: register local
35
+ provider setup and print the separate `sync-provider` command. Only an
36
+ explicit sync reads the provider, using the required admin environment
37
+ credential reference. Cursor/Copilot connectors remain fixture-verified beta
38
+ pending real-account acceptance.
39
+
40
+ ## Use standalone or connect to Workspace
41
+
42
+ Use the free CLI on its own, or keep working locally and connect it to your
43
+ invited [Tilden Workspace](https://asktilden.com/docs/workspace) for a shared
44
+ spending review. Start with `npx aibill@latest workspace connect`, complete the
45
+ browser enrollment and sharing consent, then run
46
+ `npx aibill@latest workspace push` to preview and approve a share. The browser
47
+ does not need to stay open for later pushes while the machine’s grant and
48
+ sharing consent remain valid. Shared facts include eligible
49
+ Claude Code/Codex sessions, tokens and project labels where available, plus
50
+ hashed identities; they exclude raw prompts, transcript contents, full paths,
51
+ raw session IDs and dollar amounts. Local provider records are not uploaded.
52
+ Hosted provider connections are set up separately.
53
+
54
+ Read [Workspace setup and limits](https://github.com/futurastudio/ai-spend-agent/blob/main/docs/WORKSPACE.md)
55
+ before pairing. [Join the Tilden waitlist](https://asktilden.com/?ref=npm-aibill#beta)
56
+ for updates and access invitations; joining does not grant immediate access.
57
+
58
+ ## Privacy
59
+
60
+ Default transcript analysis runs locally. Reports remain local until shared.
61
+ Explicit Workspace pushes send only the approved facts. Local MCP results go
62
+ to the AI client you configure and follow its data policy. Explicit provider
63
+ sync uses the referenced credential only with that provider’s read-only API;
64
+ aibill never sits in the inference path and never stores, prints, or proxies provider credentials.
65
+ Workspace pairing uses separate private local device credentials.
66
+
67
+ **Telemetry: command counts, disclosed at first run.** Events begin
68
+ only after the first interactive notice. They include a random installation ID
69
+ and operational fields. Events can be linked to the same installation, but do
70
+ not include your arguments, paths, prompts, project names,
71
+ dollar amounts or email. `npx aibill telemetry off` disables them; so do
72
+ `DO_NOT_TRACK`, `CI` or `AI_SPEND_NO_TELEMETRY` when non-empty.
73
+ `npx aibill telemetry` shows the exact last payload.
74
+ [Full telemetry disclosure](https://github.com/futurastudio/ai-spend-agent/blob/main/docs/TELEMETRY.md).
75
+ The optional signup command sends its displayed email/ref payload only after
76
+ confirmation; `outcome github` explicitly uses the installed `gh` CLI.
77
+
78
+ [Full guide](https://github.com/futurastudio/ai-spend-agent#readme) ·
79
+ [Source coverage](https://asktilden.com/docs/sources) ·
80
+ [Command reference](https://github.com/futurastudio/ai-spend-agent/blob/main/docs/CLI_REFERENCE.md)
package/package.json CHANGED
@@ -1,8 +1,8 @@
1
1
  {
2
2
  "name": "aibill",
3
- "version": "0.9.10",
3
+ "version": "0.9.12",
4
4
  "funding": "https://asktilden.com",
5
- "description": "Short command for ai-spend-agent: local financial accountability for Claude Code, Codex, and experimental Gemini CLI cost evidence. Run: npx aibill",
5
+ "description": "Tilden\u2019s free AI spend CLI. Inspect local Claude Code and Codex usage, keep API-rate estimates separate from provider costs. Run: npx aibill",
6
6
  "type": "module",
7
7
  "bin": {
8
8
  "aibill": "run.js"
@@ -35,6 +35,6 @@
35
35
  "node": ">=22"
36
36
  },
37
37
  "dependencies": {
38
- "ai-spend-agent": "0.9.10"
38
+ "ai-spend-agent": "0.9.12"
39
39
  }
40
40
  }