Files SDK is a unified TypeScript storage SDK for object and blob backends. It gives developers one consistent API across providers like S3, R2, GCS, and Azure Blob Storage, with a typed escape hatch for provider-specific clients.
Files SDK is a unified TypeScript storage SDK for object and blob backends. It presents one small API across providers such as AWS S3, Cloudflare R2, Google Cloud Storage, Azure Blob Storage, and other compatible services, so the same call sites can be reused when the backend changes.
The product centers on a `Files` class with standard operations like upload, download, head, list, delete, copy, move, url generation, and signed uploads. It also adds features for larger and more specialized workflows, including multipart uploads, live progress callbacks, byte-range reads, search by key, transfer and sync between instances, a CLI, and agent-oriented file tools. When provider-specific behavior is needed, the docs describe a typed `files.raw` escape hatch back to the native client.
The SDK exposes a compact `Files` surface with ten methods for common storage tasks, including bulk-friendly calls for `upload`, `delete`, and `exists`. It also adds `listAll`, `url`, `signedUploadUrl`, and a key-bound `file(key)` handle for narrower workflows.
`search()` finds objects by key using glob, regex, substring, or exact match. Results stream as an async iterable, and glob searches scope the bucket walk by prefix to avoid scanning unnecessary keys.
`upload()` supports multipart uploads for large or unbounded bodies, bounded concurrency, and configurable part size. The same operation can also report byte-level progress through `onProgress` for single or bulk uploads.
Range requests map to HTTP 206 so callers can read only part of an object. The docs call out use for video seeking, resumable downloads, and header reads without fetching the entire file.
The SDK includes ready-made file tools for Vercel AI SDK, OpenAI Agents, Claude, and MCP. It also ships a CLI that streams through stdin/stdout and defaults to JSON output.
The adapter layer covers many providers, and the core docs emphasize a typed escape hatch via `files.raw` when a backend-specific client is necessary. The API also supports `transfer` and `sync` for moving data between `Files` instances.
Use the same code path to upload, download, inspect metadata, list files, and delete objects across supported providers. This is the main fit for teams that want to switch storage backends without rewriting application logic.
Use `search()` when you need to find files by naming patterns rather than by exact key. The documented glob, regex, substring, and exact-match options make it suitable for invoice buckets, log folders, or other structured key spaces.
Use multipart uploads and `onProgress` for large files, streamed bodies, or user-facing uploads that need progress feedback. The SDK documents bounded concurrency and configurable part size for these flows.
Use byte-range downloads when you only need part of an object, such as a file header, a video segment, or a resumable chunk. The docs specifically map ranged reads to HTTP 206.
Use the CLI or agent tools when files need to be manipulated from scripts, CI jobs, or AI workflows. The source describes JSON output, stdin/stdout streaming, approval gates, and read-only modes for agent use.
It is designed for applications that need a single TypeScript storage API across object and blob backends such as S3, R2, GCS, and Azure. The docs also describe an agent-friendly CLI and a typed escape hatch to native clients when provider-specific behavior is needed.
The docs present a small unified surface: a `Files` instance with ten methods, plus `files.file(key)` for a key-bound handle. The API includes `upload`, `download`, `head`, `exists`, `delete`, `copy`, `move`, `list`, `url`, and `signedUploadUrl`, plus global `transfer` and `sync` helpers.
Yes. The adapter docs say some backends support reads and writes directly through the shared API, while provider-specific capabilities such as versioning, lifecycle rules, multipart behavior, or ACLs are available through `files.raw`, which stays typed per adapter.
The source shows a CLI that mirrors the SDK, with JSON output by default, stdin/stdout streaming, and a `--provider` flag to switch backends. It is positioned for scripts, CI, and one-off operations.
The pricing page in the collected sources is a 404 and does not state any price or plan structure, so the public site evidence here does not confirm pricing details.
Open Computer Use is an open-source Computer Use service wrapped as MCP for macOS, Linux, and Windows. It helps AI agents and MCP clients run desktop automation workflows through client setup commands or manual MCP configuration.
Lasso is an ecommerce product data platform for enriching catalog records, processing supplier files, generating product content, and monitoring competitors. It combines a web app with a REST API, SDK, and MCP server for teams and developers.
Codex Plugins bundle reusable skills, app integrations, and MCP servers into workflows you can install in the Codex app or use from Codex CLI. They help extend Codex with connected-service tasks, reusable instructions, and shared team workflows.
Struere is an AI-native platform for turning spreadsheet data into structured operational software with dashboards, alerts, and automations. It is aimed at teams that want to replace manual spreadsheet workflows without building custom tools from scratch.
Ably Chat is a chat API platform for building custom realtime chat applications. It supports room-based messaging, typing indicators, presence, reactions, and message updates, with usage-based pricing options for different deployment stages.
garden-md is an open-source Node.js CLI that turns meeting transcripts into a local company wiki, linking entities and preserving the original text.
