voltagegpu-cli 1.2.0__tar.gz → 1.3.0__tar.gz
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.
- voltagegpu_cli-1.3.0/.gitignore +14 -0
- voltagegpu_cli-1.3.0/CHANGELOG.md +87 -0
- {voltagegpu_cli-1.2.0 → voltagegpu_cli-1.3.0}/LICENSE +21 -21
- voltagegpu_cli-1.3.0/PKG-INFO +247 -0
- voltagegpu_cli-1.3.0/README.md +183 -0
- voltagegpu_cli-1.3.0/examples/deploy_standard_pod.py +52 -0
- voltagegpu_cli-1.3.0/pyproject.toml +89 -0
- voltagegpu_cli-1.3.0/tests/__init__.py +0 -0
- voltagegpu_cli-1.3.0/tests/conftest.py +130 -0
- voltagegpu_cli-1.3.0/tests/test_cli.py +243 -0
- voltagegpu_cli-1.3.0/tests/test_cli_account.py +141 -0
- voltagegpu_cli-1.3.0/tests/test_client.py +171 -0
- voltagegpu_cli-1.3.0/tests/test_config.py +90 -0
- voltagegpu_cli-1.3.0/tests/test_live.py +42 -0
- voltagegpu_cli-1.3.0/tests/test_models.py +57 -0
- voltagegpu_cli-1.3.0/tests/test_package.py +53 -0
- voltagegpu_cli-1.3.0/tests/test_ssh.py +59 -0
- voltagegpu_cli-1.3.0/volt/__init__.py +5 -0
- voltagegpu_cli-1.3.0/volt/__main__.py +5 -0
- voltagegpu_cli-1.3.0/volt/cli/__init__.py +5 -0
- voltagegpu_cli-1.3.0/volt/cli/cmd_account.py +176 -0
- voltagegpu_cli-1.3.0/volt/cli/cmd_deploy.py +192 -0
- voltagegpu_cli-1.3.0/volt/cli/cmd_machines.py +103 -0
- voltagegpu_cli-1.3.0/volt/cli/cmd_misc.py +88 -0
- voltagegpu_cli-1.3.0/volt/cli/cmd_pods.py +179 -0
- voltagegpu_cli-1.3.0/volt/cli/main.py +126 -0
- voltagegpu_cli-1.3.0/volt/cli/output.py +75 -0
- voltagegpu_cli-1.3.0/volt/cli/state.py +35 -0
- voltagegpu_cli-1.3.0/volt/sdk/__init__.py +36 -0
- voltagegpu_cli-1.3.0/volt/sdk/client.py +234 -0
- voltagegpu_cli-1.3.0/volt/sdk/config.py +185 -0
- voltagegpu_cli-1.3.0/volt/sdk/errors.py +71 -0
- voltagegpu_cli-1.3.0/volt/sdk/models.py +320 -0
- voltagegpu_cli-1.3.0/volt/sdk/ssh.py +112 -0
- voltagegpu_cli-1.2.0/.github/workflows/release.yml +0 -56
- voltagegpu_cli-1.2.0/.gitignore +0 -181
- voltagegpu_cli-1.2.0/.readthedocs.yaml +0 -15
- voltagegpu_cli-1.2.0/CHANGELOG.md +0 -61
- voltagegpu_cli-1.2.0/PKG-INFO +0 -211
- voltagegpu_cli-1.2.0/README.md +0 -174
- voltagegpu_cli-1.2.0/assets/web-app-logo.png +0 -0
- voltagegpu_cli-1.2.0/examples/deploy_confidential_pod.py +0 -60
- voltagegpu_cli-1.2.0/pyproject.toml +0 -60
- voltagegpu_cli-1.2.0/volt/__init__.py +0 -6
- voltagegpu_cli-1.2.0/volt/cli/__init__.py +0 -3
- voltagegpu_cli-1.2.0/volt/cli/cli.py +0 -476
- voltagegpu_cli-1.2.0/volt/sdk/__init__.py +0 -7
- voltagegpu_cli-1.2.0/volt/sdk/client.py +0 -215
- voltagegpu_cli-1.2.0/volt/sdk/config.py +0 -65
|
@@ -0,0 +1,87 @@
|
|
|
1
|
+
# Changelog
|
|
2
|
+
|
|
3
|
+
All notable changes to `voltagegpu-cli` are documented here. The format follows
|
|
4
|
+
[Keep a Changelog](https://keepachangelog.com/en/1.1.0/) and the project uses
|
|
5
|
+
[Semantic Versioning](https://semver.org/).
|
|
6
|
+
|
|
7
|
+
## [1.3.0] - 2026-10-10
|
|
8
|
+
|
|
9
|
+
Realigns the CLI with the product as it is today: Confidential VMs (Intel TDX)
|
|
10
|
+
and standard pods, both deployable from the public API. Every command was run
|
|
11
|
+
against the production API before release, including a paid deploy, SSH and
|
|
12
|
+
release of a standard pod.
|
|
13
|
+
|
|
14
|
+
### Added
|
|
15
|
+
- `voltagegpu` command (the `volt` command still works, same program).
|
|
16
|
+
- `voltagegpu machines [--standard | --confidential] [--gpu] [--max-price] [--min-gpus] [--json]`:
|
|
17
|
+
standard machines and Confidential VM tiers with the hourly price charged.
|
|
18
|
+
- `voltagegpu deploy standard --machine ID | --cheapest`: deploy a standard pod
|
|
19
|
+
(no enclave), with `--template`, `--ssh-key`, `--jupyter`, `--ports`,
|
|
20
|
+
`--auto-terminate`, `--wait`. Asks for confirmation before charging; scripts
|
|
21
|
+
pass `--yes`.
|
|
22
|
+
- `voltagegpu deploy confidential --tier TIER`: deploy a Confidential VM, with
|
|
23
|
+
`--hardstop`, repeatable `--port 8000[/tcp|udp]` and `--wait`. Prints the
|
|
24
|
+
one-time sudo password and the in-VM attestation commands.
|
|
25
|
+
- `voltagegpu templates`: images available for standard pods.
|
|
26
|
+
- `voltagegpu pods wait POD_ID`: wait until SSH is ready.
|
|
27
|
+
- `voltagegpu ssh POD_ID [-- COMMAND]`: run ssh directly (or `--print` it).
|
|
28
|
+
- `voltagegpu balance`: balance and running pods (`GET /api/volt/balance`).
|
|
29
|
+
- `voltagegpu login` / `logout`: store the API key in `~/.volt/config.ini`,
|
|
30
|
+
created with mode 0600, after checking that the API accepts it.
|
|
31
|
+
- `voltagegpu verify ...`: verifies a Confidential VM attestation bundle by
|
|
32
|
+
running voltage-verify (`pip install "voltagegpu-cli[verify]"`).
|
|
33
|
+
- `VOLTAGEGPU_API_KEY` is accepted as an alias of `VOLT_API_KEY`.
|
|
34
|
+
- SDK: `list_machines`, `list_templates`, `deploy_standard`,
|
|
35
|
+
`deploy_confidential`, `release_pod`, `wait_until_reachable`, `get_balance`;
|
|
36
|
+
typed immutable models; `parse_ssh_command`.
|
|
37
|
+
- Test suite (unit, CLI and opt-in live checks) and a CI workflow.
|
|
38
|
+
|
|
39
|
+
### Changed
|
|
40
|
+
- `pods stop` now releases the pod (`DELETE /api/volt/pods/:id`) and prints
|
|
41
|
+
the refunded unused prepaid time. Standard pods cannot be paused, so the old
|
|
42
|
+
`POST .../stop` call failed for them. `pods delete` and `pods get` remain as
|
|
43
|
+
aliases of `pods stop` and `pods show`.
|
|
44
|
+
- Errors show the API's own message (for example "Insufficient balance. One
|
|
45
|
+
hour costs $0.59...") instead of a raw HTTP exception, and never a traceback
|
|
46
|
+
unless `VOLT_DEBUG=1`.
|
|
47
|
+
- Python 3.10 or newer is required (3.9 reached end of life in October 2025).
|
|
48
|
+
|
|
49
|
+
### Removed
|
|
50
|
+
- `volt cc inventory`, `volt cc deploy` and `volt templates get`: Confidential
|
|
51
|
+
containers were retired on 2026-10-01. The commands now print how to deploy a
|
|
52
|
+
Confidential VM instead.
|
|
53
|
+
- `volt account info`, which called an endpoint that does not exist.
|
|
54
|
+
|
|
55
|
+
### Fixed
|
|
56
|
+
- `ssh-keys add` sent `publicKey`; the API expects `public_key`, so every key
|
|
57
|
+
upload failed with HTTP 400. It also refuses a private key file now.
|
|
58
|
+
- `account balance` called a dashboard-only endpoint and always failed with
|
|
59
|
+
HTTP 401.
|
|
60
|
+
- `cc inventory` called a dashboard-only endpoint and reported "no GPUs
|
|
61
|
+
available" instead of an error.
|
|
62
|
+
- Pod ids are validated before they are put into a URL path.
|
|
63
|
+
|
|
64
|
+
### Security
|
|
65
|
+
- The SSH command returned by the API is parsed into user, host and port, each
|
|
66
|
+
checked against a strict pattern, and ssh is started from a list of
|
|
67
|
+
arguments, never through a shell. A tampered answer such as
|
|
68
|
+
`ssh -o ProxyCommand=...` is refused.
|
|
69
|
+
- Plain `http://` base URLs are refused (except localhost), redirects are never
|
|
70
|
+
followed, and TLS certificates are always verified, so the API key cannot be
|
|
71
|
+
sent in clear or to another host.
|
|
72
|
+
- The API key is never printed (only `volt_****` and its last four
|
|
73
|
+
characters), including in `repr()` of the configuration.
|
|
74
|
+
|
|
75
|
+
## [1.2.1] - 2026-04-11
|
|
76
|
+
|
|
77
|
+
- Confidential container templates (`volt templates`), the machines list and
|
|
78
|
+
`volt cc deploy --template`.
|
|
79
|
+
|
|
80
|
+
## [1.2.0] - 2026-04-11
|
|
81
|
+
|
|
82
|
+
- `volt cc` command group for Confidential containers; default API base URL
|
|
83
|
+
`https://api.voltagegpu.com/api`; JSON output on list commands.
|
|
84
|
+
|
|
85
|
+
## [1.1.0] and [1.0.0]
|
|
86
|
+
|
|
87
|
+
- First releases of the CLI and Python SDK.
|
|
@@ -1,21 +1,21 @@
|
|
|
1
|
-
MIT License
|
|
2
|
-
|
|
3
|
-
Copyright (c)
|
|
4
|
-
|
|
5
|
-
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
-
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
-
in the Software without restriction, including without limitation the rights
|
|
8
|
-
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
-
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
-
furnished to do so, subject to the following conditions:
|
|
11
|
-
|
|
12
|
-
The above copyright notice and this permission notice shall be included in all
|
|
13
|
-
copies or substantial portions of the Software.
|
|
14
|
-
|
|
15
|
-
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
-
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
-
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
-
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
-
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
-
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
-
SOFTWARE.
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2024-2026 VOLTAGE EI (VoltageGPU)
|
|
4
|
+
|
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
+
in the Software without restriction, including without limitation the rights
|
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
+
furnished to do so, subject to the following conditions:
|
|
11
|
+
|
|
12
|
+
The above copyright notice and this permission notice shall be included in all
|
|
13
|
+
copies or substantial portions of the Software.
|
|
14
|
+
|
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
+
SOFTWARE.
|
|
@@ -0,0 +1,247 @@
|
|
|
1
|
+
Metadata-Version: 2.5
|
|
2
|
+
Name: voltagegpu-cli
|
|
3
|
+
Version: 1.3.0
|
|
4
|
+
Summary: VoltageGPU CLI and Python SDK: deploy Confidential VMs (Intel TDX) and standard GPU pods, SSH in, stop them, check your balance.
|
|
5
|
+
Project-URL: Homepage, https://voltagegpu.com
|
|
6
|
+
Project-URL: Documentation, https://docs.voltagegpu.com/cli
|
|
7
|
+
Project-URL: API reference, https://docs.voltagegpu.com/api-reference
|
|
8
|
+
Project-URL: Source, https://github.com/Jabsama/VOLTAGEGPU-CLI/tree/release/1.3
|
|
9
|
+
Project-URL: Changelog, https://github.com/Jabsama/VOLTAGEGPU-CLI/blob/release/1.3/CHANGELOG.md
|
|
10
|
+
Project-URL: Issues, https://github.com/Jabsama/VOLTAGEGPU-CLI/issues
|
|
11
|
+
Author-email: "VoltageGPU (VOLTAGE EI)" <contact@voltagegpu.com>
|
|
12
|
+
License: MIT License
|
|
13
|
+
|
|
14
|
+
Copyright (c) 2024-2026 VOLTAGE EI (VoltageGPU)
|
|
15
|
+
|
|
16
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
17
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
18
|
+
in the Software without restriction, including without limitation the rights
|
|
19
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
20
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
21
|
+
furnished to do so, subject to the following conditions:
|
|
22
|
+
|
|
23
|
+
The above copyright notice and this permission notice shall be included in all
|
|
24
|
+
copies or substantial portions of the Software.
|
|
25
|
+
|
|
26
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
27
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
28
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
29
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
30
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
31
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
32
|
+
SOFTWARE.
|
|
33
|
+
License-File: LICENSE
|
|
34
|
+
Keywords: attestation,cli,confidential-computing,gpu,intel-tdx,pytorch,sdk,ssh,voltagegpu
|
|
35
|
+
Classifier: Development Status :: 5 - Production/Stable
|
|
36
|
+
Classifier: Environment :: Console
|
|
37
|
+
Classifier: Intended Audience :: Developers
|
|
38
|
+
Classifier: Intended Audience :: Science/Research
|
|
39
|
+
Classifier: License :: OSI Approved :: MIT License
|
|
40
|
+
Classifier: Operating System :: OS Independent
|
|
41
|
+
Classifier: Programming Language :: Python :: 3
|
|
42
|
+
Classifier: Programming Language :: Python :: 3 :: Only
|
|
43
|
+
Classifier: Programming Language :: Python :: 3.10
|
|
44
|
+
Classifier: Programming Language :: Python :: 3.11
|
|
45
|
+
Classifier: Programming Language :: Python :: 3.12
|
|
46
|
+
Classifier: Programming Language :: Python :: 3.13
|
|
47
|
+
Classifier: Programming Language :: Python :: 3.14
|
|
48
|
+
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
|
|
49
|
+
Classifier: Topic :: Security
|
|
50
|
+
Classifier: Topic :: System :: Systems Administration
|
|
51
|
+
Requires-Python: >=3.10
|
|
52
|
+
Requires-Dist: click>=8.1
|
|
53
|
+
Requires-Dist: httpx<1,>=0.27
|
|
54
|
+
Requires-Dist: rich>=13.0
|
|
55
|
+
Provides-Extra: dev
|
|
56
|
+
Requires-Dist: build>=1.2; extra == 'dev'
|
|
57
|
+
Requires-Dist: pytest-cov>=5; extra == 'dev'
|
|
58
|
+
Requires-Dist: pytest>=8; extra == 'dev'
|
|
59
|
+
Requires-Dist: ruff>=0.6; extra == 'dev'
|
|
60
|
+
Requires-Dist: twine>=5; extra == 'dev'
|
|
61
|
+
Provides-Extra: verify
|
|
62
|
+
Requires-Dist: voltage-verify>=0.2.1; extra == 'verify'
|
|
63
|
+
Description-Content-Type: text/markdown
|
|
64
|
+
|
|
65
|
+
# voltagegpu-cli
|
|
66
|
+
|
|
67
|
+
Deploy and manage VoltageGPU machines from your terminal or from Python:
|
|
68
|
+
|
|
69
|
+
- **Confidential VMs**: Intel TDX trust domains with an NVIDIA GPU, only your SSH
|
|
70
|
+
keys on the machine, and an attestation you generate and verify yourself.
|
|
71
|
+
- **Standard pods**: whole GPU machines without an enclave, for workloads that
|
|
72
|
+
are not sensitive.
|
|
73
|
+
|
|
74
|
+
Billing is per second. The first hour is charged when you deploy; when you stop
|
|
75
|
+
a standard pod, the unused part of that hour goes back to your balance.
|
|
76
|
+
|
|
77
|
+
```bash
|
|
78
|
+
pip install voltagegpu-cli
|
|
79
|
+
```
|
|
80
|
+
|
|
81
|
+
Python 3.10 or newer. The package installs the `voltagegpu` command (`volt` is
|
|
82
|
+
an alias kept from earlier releases).
|
|
83
|
+
|
|
84
|
+
## 1. Log in
|
|
85
|
+
|
|
86
|
+
Create an API key at <https://voltagegpu.com/api-keys> (it is shown once), then:
|
|
87
|
+
|
|
88
|
+
```bash
|
|
89
|
+
voltagegpu login # prompts for the key, hidden, checks it, saves it to ~/.volt/config.ini (mode 0600)
|
|
90
|
+
# or, for CI and scripts:
|
|
91
|
+
export VOLT_API_KEY=volt_...
|
|
92
|
+
```
|
|
93
|
+
|
|
94
|
+
Add the SSH key you will connect with (defaults to `~/.ssh/id_ed25519.pub`):
|
|
95
|
+
|
|
96
|
+
```bash
|
|
97
|
+
voltagegpu ssh-keys add
|
|
98
|
+
voltagegpu ssh-keys list
|
|
99
|
+
```
|
|
100
|
+
|
|
101
|
+
## 2. Pick a machine
|
|
102
|
+
|
|
103
|
+
```bash
|
|
104
|
+
voltagegpu machines # Confidential VM tiers, then standard machines, cheapest first
|
|
105
|
+
voltagegpu machines --standard --gpu A6000 # filter by GPU model
|
|
106
|
+
voltagegpu machines --standard --max-price 1 --json
|
|
107
|
+
voltagegpu machines --confidential
|
|
108
|
+
voltagegpu templates # images for standard pods (PyTorch by default)
|
|
109
|
+
```
|
|
110
|
+
|
|
111
|
+
## 3a. Deploy a standard pod
|
|
112
|
+
|
|
113
|
+
```bash
|
|
114
|
+
voltagegpu deploy standard --machine <machine_id>
|
|
115
|
+
voltagegpu deploy standard --cheapest --gpu 4090 --wait # cheapest matching machine, wait for SSH
|
|
116
|
+
```
|
|
117
|
+
|
|
118
|
+
The CLI shows the price and asks before charging. In scripts, add `--yes`.
|
|
119
|
+
If your account has several SSH keys, choose one with `--ssh-key <key_id>`.
|
|
120
|
+
Other options: `--template <template_id>`, `--jupyter`, `--ports 3`,
|
|
121
|
+
`--auto-terminate 8` (hours), `--name my-pod`.
|
|
122
|
+
|
|
123
|
+
## 3b. Deploy a Confidential VM
|
|
124
|
+
|
|
125
|
+
```bash
|
|
126
|
+
voltagegpu deploy confidential --tier rtx6000b-small --wait
|
|
127
|
+
voltagegpu deploy confidential --tier h100-small --hardstop 4 --port 8000 --port 8443/tcp
|
|
128
|
+
```
|
|
129
|
+
|
|
130
|
+
Every SSH key on your account is bound to the VM when it is created; a VM
|
|
131
|
+
cannot be given a key afterwards, so add yours first. The sudo password is
|
|
132
|
+
printed once and is not kept anywhere. The command also prints how to generate
|
|
133
|
+
the attestation inside the VM.
|
|
134
|
+
|
|
135
|
+
## 4. Connect, inspect, stop
|
|
136
|
+
|
|
137
|
+
```bash
|
|
138
|
+
voltagegpu pods list
|
|
139
|
+
voltagegpu pods wait <pod_id> # until SSH answers
|
|
140
|
+
voltagegpu ssh <pod_id> # interactive shell
|
|
141
|
+
voltagegpu ssh <pod_id> -- nvidia-smi -L # run one command
|
|
142
|
+
voltagegpu ssh <pod_id> --print # just print the ssh command
|
|
143
|
+
voltagegpu pods show <pod_id>
|
|
144
|
+
voltagegpu pods stop <pod_id> # releases it: billing stops, the disk is deleted
|
|
145
|
+
voltagegpu balance
|
|
146
|
+
```
|
|
147
|
+
|
|
148
|
+
Fresh pods present a new host key. In non-interactive scripts you can accept it
|
|
149
|
+
on first use with `-o StrictHostKeyChecking=accept-new`.
|
|
150
|
+
|
|
151
|
+
## 5. Verify a Confidential VM
|
|
152
|
+
|
|
153
|
+
Inside the VM you produce an evidence bundle (the deploy output prints the exact
|
|
154
|
+
commands). On your own machine:
|
|
155
|
+
|
|
156
|
+
```bash
|
|
157
|
+
pip install "voltagegpu-cli[verify]" # installs voltage-verify
|
|
158
|
+
voltagegpu verify bundle.json --offline
|
|
159
|
+
voltagegpu verify bundle.json --challenge <hex-you-issued> --gpus 1
|
|
160
|
+
```
|
|
161
|
+
|
|
162
|
+
`voltagegpu verify` runs [voltage-verify](https://pypi.org/project/voltage-verify/),
|
|
163
|
+
which checks the Intel TDX quote up to Intel's root and the NVIDIA GPU
|
|
164
|
+
attestation tokens, without trusting VoltageGPU.
|
|
165
|
+
|
|
166
|
+
## Scripting
|
|
167
|
+
|
|
168
|
+
Every list command takes `--json`; deploys and stops take `--yes`.
|
|
169
|
+
|
|
170
|
+
```bash
|
|
171
|
+
# Deploy the cheapest standard machine, wait for SSH, run a job, release it.
|
|
172
|
+
POD=$(voltagegpu deploy standard --cheapest --yes --json | jq -r .deploy.pod_id)
|
|
173
|
+
voltagegpu pods wait "$POD"
|
|
174
|
+
voltagegpu ssh "$POD" -o StrictHostKeyChecking=accept-new -- nvidia-smi
|
|
175
|
+
voltagegpu pods stop "$POD" --yes
|
|
176
|
+
```
|
|
177
|
+
|
|
178
|
+
Exit codes: `0` success, `1` API or runtime error, `2` usage error or missing
|
|
179
|
+
confirmation.
|
|
180
|
+
|
|
181
|
+
## Python SDK
|
|
182
|
+
|
|
183
|
+
```python
|
|
184
|
+
from volt.sdk import VoltageGPUClient
|
|
185
|
+
|
|
186
|
+
with VoltageGPUClient() as client: # VOLT_API_KEY or ~/.volt/config.ini
|
|
187
|
+
machine = client.list_machines("standard")[0] # cheapest first
|
|
188
|
+
result = client.deploy_standard(machine.id, name="my-pod")
|
|
189
|
+
pod = client.wait_until_reachable(result.pod_id)
|
|
190
|
+
print(pod.ssh_command)
|
|
191
|
+
print(client.release_pod(result.pod_id).prepaid_refund)
|
|
192
|
+
```
|
|
193
|
+
|
|
194
|
+
A complete example is in [`examples/deploy_standard_pod.py`](examples/deploy_standard_pod.py).
|
|
195
|
+
|
|
196
|
+
## REST API
|
|
197
|
+
|
|
198
|
+
The CLI is a thin client of the public API, documented at
|
|
199
|
+
<https://docs.voltagegpu.com/api-reference>. The same deploy with curl:
|
|
200
|
+
|
|
201
|
+
```bash
|
|
202
|
+
curl -X POST https://api.voltagegpu.com/api/volt/pods \
|
|
203
|
+
-H "X-API-Key: $VOLT_API_KEY" -H "Content-Type: application/json" \
|
|
204
|
+
-d '{"provider":"standard","name":"my-pod","machine_id":"<machine_id>"}'
|
|
205
|
+
|
|
206
|
+
curl https://api.voltagegpu.com/api/volt/pods -H "X-API-Key: $VOLT_API_KEY"
|
|
207
|
+
```
|
|
208
|
+
|
|
209
|
+
## Configuration
|
|
210
|
+
|
|
211
|
+
| Setting | Where |
|
|
212
|
+
| --- | --- |
|
|
213
|
+
| API key | `--api-key` of `login`, `VOLT_API_KEY` (or `VOLTAGEGPU_API_KEY`), then `~/.volt/config.ini` |
|
|
214
|
+
| API base URL | `VOLT_BASE_URL` (https only; default `https://api.voltagegpu.com/api`) |
|
|
215
|
+
| Debug | `VOLT_DEBUG=1` shows tracebacks |
|
|
216
|
+
|
|
217
|
+
`voltagegpu config` shows where the key comes from, masked.
|
|
218
|
+
|
|
219
|
+
## Security
|
|
220
|
+
|
|
221
|
+
- The key is sent only over HTTPS, in the `X-API-Key` header; redirects are
|
|
222
|
+
never followed and certificates are always verified.
|
|
223
|
+
- The key is never printed. `~/.volt/config.ini` is created readable by you
|
|
224
|
+
only, and `voltagegpu config` warns if its permissions are too open.
|
|
225
|
+
- The SSH command returned by the API is parsed into user, host and port and
|
|
226
|
+
ssh is started without a shell, so a tampered answer cannot run a command on
|
|
227
|
+
your machine.
|
|
228
|
+
- `ssh-keys add` refuses private keys.
|
|
229
|
+
|
|
230
|
+
Report a vulnerability to contact@voltagegpu.com.
|
|
231
|
+
|
|
232
|
+
## Upgrading from 1.2
|
|
233
|
+
|
|
234
|
+
- `volt cc inventory` / `volt cc deploy`: Confidential containers were retired
|
|
235
|
+
on 2026-10-01. Use `voltagegpu machines --confidential` and
|
|
236
|
+
`voltagegpu deploy confidential --tier <tier>`.
|
|
237
|
+
- `volt pods stop` now releases the pod and reports the refund.
|
|
238
|
+
- `volt account balance` is now `voltagegpu balance`.
|
|
239
|
+
|
|
240
|
+
## Links
|
|
241
|
+
|
|
242
|
+
- Docs: <https://docs.voltagegpu.com/cli>
|
|
243
|
+
- API reference: <https://docs.voltagegpu.com/api-reference>
|
|
244
|
+
- API keys: <https://voltagegpu.com/api-keys>
|
|
245
|
+
- Changelog: [CHANGELOG.md](CHANGELOG.md)
|
|
246
|
+
|
|
247
|
+
MIT License.
|
|
@@ -0,0 +1,183 @@
|
|
|
1
|
+
# voltagegpu-cli
|
|
2
|
+
|
|
3
|
+
Deploy and manage VoltageGPU machines from your terminal or from Python:
|
|
4
|
+
|
|
5
|
+
- **Confidential VMs**: Intel TDX trust domains with an NVIDIA GPU, only your SSH
|
|
6
|
+
keys on the machine, and an attestation you generate and verify yourself.
|
|
7
|
+
- **Standard pods**: whole GPU machines without an enclave, for workloads that
|
|
8
|
+
are not sensitive.
|
|
9
|
+
|
|
10
|
+
Billing is per second. The first hour is charged when you deploy; when you stop
|
|
11
|
+
a standard pod, the unused part of that hour goes back to your balance.
|
|
12
|
+
|
|
13
|
+
```bash
|
|
14
|
+
pip install voltagegpu-cli
|
|
15
|
+
```
|
|
16
|
+
|
|
17
|
+
Python 3.10 or newer. The package installs the `voltagegpu` command (`volt` is
|
|
18
|
+
an alias kept from earlier releases).
|
|
19
|
+
|
|
20
|
+
## 1. Log in
|
|
21
|
+
|
|
22
|
+
Create an API key at <https://voltagegpu.com/api-keys> (it is shown once), then:
|
|
23
|
+
|
|
24
|
+
```bash
|
|
25
|
+
voltagegpu login # prompts for the key, hidden, checks it, saves it to ~/.volt/config.ini (mode 0600)
|
|
26
|
+
# or, for CI and scripts:
|
|
27
|
+
export VOLT_API_KEY=volt_...
|
|
28
|
+
```
|
|
29
|
+
|
|
30
|
+
Add the SSH key you will connect with (defaults to `~/.ssh/id_ed25519.pub`):
|
|
31
|
+
|
|
32
|
+
```bash
|
|
33
|
+
voltagegpu ssh-keys add
|
|
34
|
+
voltagegpu ssh-keys list
|
|
35
|
+
```
|
|
36
|
+
|
|
37
|
+
## 2. Pick a machine
|
|
38
|
+
|
|
39
|
+
```bash
|
|
40
|
+
voltagegpu machines # Confidential VM tiers, then standard machines, cheapest first
|
|
41
|
+
voltagegpu machines --standard --gpu A6000 # filter by GPU model
|
|
42
|
+
voltagegpu machines --standard --max-price 1 --json
|
|
43
|
+
voltagegpu machines --confidential
|
|
44
|
+
voltagegpu templates # images for standard pods (PyTorch by default)
|
|
45
|
+
```
|
|
46
|
+
|
|
47
|
+
## 3a. Deploy a standard pod
|
|
48
|
+
|
|
49
|
+
```bash
|
|
50
|
+
voltagegpu deploy standard --machine <machine_id>
|
|
51
|
+
voltagegpu deploy standard --cheapest --gpu 4090 --wait # cheapest matching machine, wait for SSH
|
|
52
|
+
```
|
|
53
|
+
|
|
54
|
+
The CLI shows the price and asks before charging. In scripts, add `--yes`.
|
|
55
|
+
If your account has several SSH keys, choose one with `--ssh-key <key_id>`.
|
|
56
|
+
Other options: `--template <template_id>`, `--jupyter`, `--ports 3`,
|
|
57
|
+
`--auto-terminate 8` (hours), `--name my-pod`.
|
|
58
|
+
|
|
59
|
+
## 3b. Deploy a Confidential VM
|
|
60
|
+
|
|
61
|
+
```bash
|
|
62
|
+
voltagegpu deploy confidential --tier rtx6000b-small --wait
|
|
63
|
+
voltagegpu deploy confidential --tier h100-small --hardstop 4 --port 8000 --port 8443/tcp
|
|
64
|
+
```
|
|
65
|
+
|
|
66
|
+
Every SSH key on your account is bound to the VM when it is created; a VM
|
|
67
|
+
cannot be given a key afterwards, so add yours first. The sudo password is
|
|
68
|
+
printed once and is not kept anywhere. The command also prints how to generate
|
|
69
|
+
the attestation inside the VM.
|
|
70
|
+
|
|
71
|
+
## 4. Connect, inspect, stop
|
|
72
|
+
|
|
73
|
+
```bash
|
|
74
|
+
voltagegpu pods list
|
|
75
|
+
voltagegpu pods wait <pod_id> # until SSH answers
|
|
76
|
+
voltagegpu ssh <pod_id> # interactive shell
|
|
77
|
+
voltagegpu ssh <pod_id> -- nvidia-smi -L # run one command
|
|
78
|
+
voltagegpu ssh <pod_id> --print # just print the ssh command
|
|
79
|
+
voltagegpu pods show <pod_id>
|
|
80
|
+
voltagegpu pods stop <pod_id> # releases it: billing stops, the disk is deleted
|
|
81
|
+
voltagegpu balance
|
|
82
|
+
```
|
|
83
|
+
|
|
84
|
+
Fresh pods present a new host key. In non-interactive scripts you can accept it
|
|
85
|
+
on first use with `-o StrictHostKeyChecking=accept-new`.
|
|
86
|
+
|
|
87
|
+
## 5. Verify a Confidential VM
|
|
88
|
+
|
|
89
|
+
Inside the VM you produce an evidence bundle (the deploy output prints the exact
|
|
90
|
+
commands). On your own machine:
|
|
91
|
+
|
|
92
|
+
```bash
|
|
93
|
+
pip install "voltagegpu-cli[verify]" # installs voltage-verify
|
|
94
|
+
voltagegpu verify bundle.json --offline
|
|
95
|
+
voltagegpu verify bundle.json --challenge <hex-you-issued> --gpus 1
|
|
96
|
+
```
|
|
97
|
+
|
|
98
|
+
`voltagegpu verify` runs [voltage-verify](https://pypi.org/project/voltage-verify/),
|
|
99
|
+
which checks the Intel TDX quote up to Intel's root and the NVIDIA GPU
|
|
100
|
+
attestation tokens, without trusting VoltageGPU.
|
|
101
|
+
|
|
102
|
+
## Scripting
|
|
103
|
+
|
|
104
|
+
Every list command takes `--json`; deploys and stops take `--yes`.
|
|
105
|
+
|
|
106
|
+
```bash
|
|
107
|
+
# Deploy the cheapest standard machine, wait for SSH, run a job, release it.
|
|
108
|
+
POD=$(voltagegpu deploy standard --cheapest --yes --json | jq -r .deploy.pod_id)
|
|
109
|
+
voltagegpu pods wait "$POD"
|
|
110
|
+
voltagegpu ssh "$POD" -o StrictHostKeyChecking=accept-new -- nvidia-smi
|
|
111
|
+
voltagegpu pods stop "$POD" --yes
|
|
112
|
+
```
|
|
113
|
+
|
|
114
|
+
Exit codes: `0` success, `1` API or runtime error, `2` usage error or missing
|
|
115
|
+
confirmation.
|
|
116
|
+
|
|
117
|
+
## Python SDK
|
|
118
|
+
|
|
119
|
+
```python
|
|
120
|
+
from volt.sdk import VoltageGPUClient
|
|
121
|
+
|
|
122
|
+
with VoltageGPUClient() as client: # VOLT_API_KEY or ~/.volt/config.ini
|
|
123
|
+
machine = client.list_machines("standard")[0] # cheapest first
|
|
124
|
+
result = client.deploy_standard(machine.id, name="my-pod")
|
|
125
|
+
pod = client.wait_until_reachable(result.pod_id)
|
|
126
|
+
print(pod.ssh_command)
|
|
127
|
+
print(client.release_pod(result.pod_id).prepaid_refund)
|
|
128
|
+
```
|
|
129
|
+
|
|
130
|
+
A complete example is in [`examples/deploy_standard_pod.py`](examples/deploy_standard_pod.py).
|
|
131
|
+
|
|
132
|
+
## REST API
|
|
133
|
+
|
|
134
|
+
The CLI is a thin client of the public API, documented at
|
|
135
|
+
<https://docs.voltagegpu.com/api-reference>. The same deploy with curl:
|
|
136
|
+
|
|
137
|
+
```bash
|
|
138
|
+
curl -X POST https://api.voltagegpu.com/api/volt/pods \
|
|
139
|
+
-H "X-API-Key: $VOLT_API_KEY" -H "Content-Type: application/json" \
|
|
140
|
+
-d '{"provider":"standard","name":"my-pod","machine_id":"<machine_id>"}'
|
|
141
|
+
|
|
142
|
+
curl https://api.voltagegpu.com/api/volt/pods -H "X-API-Key: $VOLT_API_KEY"
|
|
143
|
+
```
|
|
144
|
+
|
|
145
|
+
## Configuration
|
|
146
|
+
|
|
147
|
+
| Setting | Where |
|
|
148
|
+
| --- | --- |
|
|
149
|
+
| API key | `--api-key` of `login`, `VOLT_API_KEY` (or `VOLTAGEGPU_API_KEY`), then `~/.volt/config.ini` |
|
|
150
|
+
| API base URL | `VOLT_BASE_URL` (https only; default `https://api.voltagegpu.com/api`) |
|
|
151
|
+
| Debug | `VOLT_DEBUG=1` shows tracebacks |
|
|
152
|
+
|
|
153
|
+
`voltagegpu config` shows where the key comes from, masked.
|
|
154
|
+
|
|
155
|
+
## Security
|
|
156
|
+
|
|
157
|
+
- The key is sent only over HTTPS, in the `X-API-Key` header; redirects are
|
|
158
|
+
never followed and certificates are always verified.
|
|
159
|
+
- The key is never printed. `~/.volt/config.ini` is created readable by you
|
|
160
|
+
only, and `voltagegpu config` warns if its permissions are too open.
|
|
161
|
+
- The SSH command returned by the API is parsed into user, host and port and
|
|
162
|
+
ssh is started without a shell, so a tampered answer cannot run a command on
|
|
163
|
+
your machine.
|
|
164
|
+
- `ssh-keys add` refuses private keys.
|
|
165
|
+
|
|
166
|
+
Report a vulnerability to contact@voltagegpu.com.
|
|
167
|
+
|
|
168
|
+
## Upgrading from 1.2
|
|
169
|
+
|
|
170
|
+
- `volt cc inventory` / `volt cc deploy`: Confidential containers were retired
|
|
171
|
+
on 2026-10-01. Use `voltagegpu machines --confidential` and
|
|
172
|
+
`voltagegpu deploy confidential --tier <tier>`.
|
|
173
|
+
- `volt pods stop` now releases the pod and reports the refund.
|
|
174
|
+
- `volt account balance` is now `voltagegpu balance`.
|
|
175
|
+
|
|
176
|
+
## Links
|
|
177
|
+
|
|
178
|
+
- Docs: <https://docs.voltagegpu.com/cli>
|
|
179
|
+
- API reference: <https://docs.voltagegpu.com/api-reference>
|
|
180
|
+
- API keys: <https://voltagegpu.com/api-keys>
|
|
181
|
+
- Changelog: [CHANGELOG.md](CHANGELOG.md)
|
|
182
|
+
|
|
183
|
+
MIT License.
|
|
@@ -0,0 +1,52 @@
|
|
|
1
|
+
"""Deploy the cheapest standard pod, wait for SSH, run nvidia-smi, release it.
|
|
2
|
+
|
|
3
|
+
export VOLT_API_KEY=volt_...
|
|
4
|
+
python examples/deploy_standard_pod.py --max-price 1.0
|
|
5
|
+
|
|
6
|
+
This charges your balance: one hour up front, refunded down to the seconds used
|
|
7
|
+
when the pod is released (the script always releases it, even on error).
|
|
8
|
+
"""
|
|
9
|
+
|
|
10
|
+
from __future__ import annotations
|
|
11
|
+
|
|
12
|
+
import argparse
|
|
13
|
+
import shutil
|
|
14
|
+
import subprocess
|
|
15
|
+
import sys
|
|
16
|
+
|
|
17
|
+
from volt.sdk import ApiError, VoltageGPUClient, parse_ssh_command
|
|
18
|
+
|
|
19
|
+
|
|
20
|
+
def main() -> int:
|
|
21
|
+
parser = argparse.ArgumentParser(description=__doc__.splitlines()[0])
|
|
22
|
+
parser.add_argument("--max-price", type=float, default=1.0, help="maximum price per hour (USD)")
|
|
23
|
+
parser.add_argument("--ssh-key", help="SSH key id, if your account has several")
|
|
24
|
+
args = parser.parse_args()
|
|
25
|
+
|
|
26
|
+
with VoltageGPUClient() as client:
|
|
27
|
+
machines = [m for m in client.list_machines("standard") if m.price_per_hour and m.price_per_hour <= args.max_price]
|
|
28
|
+
if not machines:
|
|
29
|
+
print(f"No standard machine under ${args.max_price}/hour right now.")
|
|
30
|
+
return 1
|
|
31
|
+
machine = machines[0]
|
|
32
|
+
print(f"Deploying on {machine.gpu_count}x {machine.gpu_type} at ${machine.price_per_hour}/hour")
|
|
33
|
+
result = client.deploy_standard(machine.id, name="sdk-example", ssh_key_id=args.ssh_key)
|
|
34
|
+
try:
|
|
35
|
+
pod = client.wait_until_reachable(result.pod_id, timeout=600)
|
|
36
|
+
print("SSH:", pod.ssh_command)
|
|
37
|
+
if shutil.which("ssh") and pod.ssh_command:
|
|
38
|
+
argv = parse_ssh_command(pod.ssh_command).argv(
|
|
39
|
+
options=["StrictHostKeyChecking=accept-new", "ConnectTimeout=15"],
|
|
40
|
+
remote_command=["nvidia-smi", "-L"],
|
|
41
|
+
)
|
|
42
|
+
subprocess.call(argv) # noqa: S603 (argv rebuilt from validated user, host, port)
|
|
43
|
+
except (ApiError, TimeoutError) as exc:
|
|
44
|
+
print("Pod did not become reachable:", exc)
|
|
45
|
+
finally:
|
|
46
|
+
released = client.release_pod(result.pod_id)
|
|
47
|
+
print(f"Released {result.pod_id}, refunded ${released.prepaid_refund or 0:.4f}")
|
|
48
|
+
return 0
|
|
49
|
+
|
|
50
|
+
|
|
51
|
+
if __name__ == "__main__":
|
|
52
|
+
sys.exit(main())
|