canonical: https://jentic.com/apis/api.paperspace.com/paperspace

# Paperspace API

Jentic publishes the only available OpenAPI specification for Paperspace API, keeping it validated and agent-ready. Paperspace provides GPU-backed compute and machine learning infrastructure, and the API exposes the resources teams use to organise that work - projects, deployments, project collaborators, secrets, and tags. Authenticated callers can list and create projects, manage who can access them, store project-level secrets, and inspect activity. Bearer authentication is used throughout, with `/auth/session` as the entry point for token validation.

## For AI agents

Provision and manage Paperspace projects, deployments, collaborators, and secrets through bearer-authenticated calls - useful for orchestrating GPU-backed ML workloads.

## Scope

Does not run model training jobs directly, manage billing, or expose notebook content - use for project, deployment, collaborator, and secret management only.

## Capabilities

- List and create Paperspace projects via /projects and inspect a single project at `/projects/{id}`
- Add or remove collaborators on a project through `/projects/{id}/collaborators` and `/projects/{id}/collaborators/{userId}`
- Store and retrieve project-level secrets with `/projects/{id}/secrets` and `/projects/{id}/secrets/{name}`
- Tag projects for organisation through `/projects/{id}/tags`
- Inspect recent project activity via `/projects/{id}/activity`
- Manage deployments through the /deployments resource
- Validate the active session and bearer token at `/auth/session`

## Use cases

### Project Provisioning Automation

Stand up new Paperspace projects on demand by calling /projects with a name and description, then attach collaborators and secrets in follow-up calls. This supports onboarding flows that give each new ML engineer their own workspace, or per-experiment provisioning where each run lives in an isolated project. Tags and activity endpoints help downstream tooling track which projects are active.

Example prompt: POST /projects with name='exp-2026-06-09', then POST `/projects/{id}/tags` with 'experiment' and POST `/projects/{id}/secrets` to store the run's OPENAI_API_KEY.

### Collaborator Access Control

Manage who can access each Paperspace project through `/projects/{id}/collaborators` and the corresponding DELETE on `/projects/{id}/collaborators/{userId}.` This keeps access tightly scoped - a teammate can be added for a single experiment and removed once it ships, without touching every other project. Combined with /activity, audit trails are easy to build.

Example prompt: List collaborators on a project, add a new user by id, then remove a previous collaborator whose access has expired.

### Centralised Secret Storage

Store API keys, tokens, and configuration as project secrets via `/projects/{id}/secrets` so notebooks and deployments can read them at runtime without checking values into the repository. `/projects/{id}/secrets/{name}` supports retrieval and deletion of individual secrets. Pair with the bearer-authenticated session to keep operational credentials out of source control.

Example prompt: POST `/projects/{id}/secrets` with name and value, then GET `/projects/{id}/secrets` to confirm the secret is registered.

### AI Agent Workspace Orchestration via Jentic

Have an agent provision a Paperspace project, attach collaborators, and seed secrets in response to a higher-level instruction such as 'spin up an experiment workspace for me and Sam'. Jentic returns the right operation per intent, the agent supplies parameters, and the bearer token stays in your Jentic One instance - keeping ML infrastructure credentials out of agent prompts.

Example prompt: Search Jentic for 'create a Paperspace project', load the POST /projects schema, execute with the requested name, then chain calls to add collaborators and store an initial secret.

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| GET | `/projects` | List projects |
| POST | `/projects` | Create a project |
| GET | `/projects/{id}` | Retrieve a project |
| GET | `/projects/{id}/collaborators` | List project collaborators |
| GET | `/projects/{id}/secrets` | List project secrets |
| GET | `/projects/{id}/activity` | Inspect project activity |
| GET | `/deployments` | List deployments |
| GET | `/auth/session` | Validate the current session |

## Key resources

- **Projects** — Create, list, and manage projects
- **Collaborators** — Add and remove project collaborators
- **Secrets** — Project-scoped secret storage
- **Tags** — Tag projects for organisation
- **Activity** — Inspect recent project activity
- **Deployments** — Manage Paperspace deployments
- **Auth** — Validate and inspect the current session

## Why Jentic

- **Setup:** Wiring Paperspace by hand means setting up its bearer auth, learning the project, deployment, and secret resource shapes, and managing collaborators yourself. Through Jentic you install once, import Paperspace from the API Directory, store the token once, and your agent calls it.
- **Permission scoping:** Paperspace puts the project id in the URL path (`/projects/{id}/...`), so a rule can pin your agent to one project: it can read that project's collaborators, secrets, and activity and nothing else. You choose the operations it may call, so creating projects is not included unless you add it.
- **Credential handling:** Your Paperspace token is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'list deployments' or 'read a project's secrets', and Jentic returns the matching Paperspace operation with its input schema so the agent calls the right endpoint without browsing the reference docs.

## Related APIs

- **Replicate API** — Run open-source ML models on managed GPU infrastructure
- **Hugging Face API** — Models, datasets, and inference endpoints
- **DigitalOcean API** — General-purpose cloud compute and storage

## FAQ

### Why is there no official OpenAPI spec for Paperspace API?

Paperspace does not publish an OpenAPI specification. Jentic generates and maintains this spec so that AI agents and developers can call Paperspace API via structured tooling. It is validated against the live API and kept up to date. Get started with Jentic One, the self-hosted execution layer.

### What authentication does the Paperspace API use?

The Paperspace API uses HTTP bearer token authentication. The token is sent in the Authorization header on every request, and `/auth/session` can be called to verify it. Through Jentic, the bearer token is stored encrypted in your Jentic One instance and injected at execution time, so it never appears in agent prompts.

### Can I manage project secrets with the Paperspace API?

Yes. POST `/projects/{id}/secrets` stores a named secret on a project, GET `/projects/{id}/secrets/{name}` retrieves it, and DELETE on the same path removes it. This keeps third-party API keys (e.g., model provider tokens) out of source control while still being available to notebooks and deployments.

### What are the rate limits for the Paperspace API?

The published spec does not declare numeric rate limits. Treat /projects and /deployments as the highest-volume endpoints, paginate listings, and back off on HTTP 429. Validation calls to `/auth/session` should be cached rather than issued on every request.

### How do I create a Paperspace project through Jentic?

Search Jentic for 'create a Paperspace project', load the POST /projects schema, and execute with the project name and description. With `pip install jentic` the call returns the new project id, which you can pass to follow-up calls for collaborators, secrets, or tags.

### Is the Paperspace API free?

API access itself is free; the underlying compute (GPU machines, deployments) is billed according to Paperspace's pricing. Project, collaborator, secret, and tag operations consume only standard API quota.

### Can I limit what my agent is allowed to do with the Paperspace API?

Yes. Because Jentic One is self-hosted, you set the rules that decide which Paperspace operations your agent may call and which stored token it uses. Since Paperspace puts the project id in the URL path (`/projects/{id}/...`), you can pin the agent to a single project so it only reads that project's collaborators, secrets, and activity and touches nothing else. You choose the allowed operations, so an action like creating new projects with POST /projects stays off unless you explicitly grant it.