Agents
Where do AI agents save their work? 24 artifact stores compared
An agent makes things. It writes a report, fills a table, builds a ledger or saves a file. The next session has to find that work and change it without breaking it. This post is about where that work lives. It is not about memory, the facts and preferences an agent remembers about a person.
We read the docs of 23 places an agent can save its work on 16 November 2026, and set our own beside them. For each one we asked four questions. Does it keep old versions? Does it refuse a write made from an old copy? Does it check what is written? Can another agent or a person read the same thing?
5 of 23
keep earlier versions of what the agent saved
2 of 23
refuse a write made from an out-of-date copy
3 of 23
check what is written against a type or a schema
The 24 stores
Each row says only what the maker’s own page says. “Not documented” means the page we read does not say either way. Our own store, hedwigai workbooks, is the last row.
| Store | Maker | What the agent saves | Old versions kept | Stale write refused | Checked on write | Who else can read it | Access |
|---|---|---|---|---|---|---|---|
| Agent Bucket | Tencent Cloud | Files in a Space, one per agent, user or team | Not documented | Not documented | No | Shared Spaces and share links | Console, REST API, COS SDK |
| ADK ArtifactService | Named files with a MIME type | Yes, a number for every save | Not documented | No | The same user's other sessions, with the user: prefix | ADK SDK | |
| AgentCore session storage | AWS | A directory at a mount path, reset after 14 idle days | No | No | No | No: each session sees only its own | Runtime configuration |
| Files API | Anthropic | Uploads, and files made by code execution and skills | No: a file cannot change, so a new one replaces it | Not needed: files cannot change | No | Anyone with a key to the workspace | API |
| Managed Agents session outputs | Anthropic | Files the agent writes to its outputs folder | No | Not documented | No | Through the Files API | API |
| Managed Agents memory stores (beta) | Anthropic | Text documents at paths | Yes, each change kept 30 days | Yes, by a content hash | No | Other sessions, read-only or read-write | API, a mounted folder |
| Containers (Code Interpreter) | OpenAI | Files in a container that expires after 20 idle minutes | Not documented | Not documented | No | Not documented | API |
| Conversations API | OpenAI | Conversation items, tool outputs among them | No | Not documented | No | Not documented | API |
| Agents SDK Sessions | OpenAI | Chat history in SQLite, Redis and other databases | Branching in one session type | Not documented | No | Not documented | SDK |
| Foundry Agent Service | Microsoft | Files in your Azure Storage, threads in Cosmos DB | Not documented | Not documented | No | Inside your own tenant | API, SDK |
| LangGraph checkpointers | LangChain | The graph's state, per thread | Yes, the checkpoint history | Not documented | Not documented | The same thread | SDK |
| LangGraph Store | LangChain | Key-value items in namespaces | Not documented | Not documented | Not documented | Other threads | SDK |
| Flows @persist | CrewAI | A flow's state in SQLite | Snapshots, to resume or fork | Not documented | Yes, Pydantic types on the state | Not documented | SDK |
| Agents SDK state | Cloudflare | JSON state and SQL tables in a Durable Object | Not documented | Not documented | Yes, a validateStateChange hook | Connected clients, kept in sync | SDK |
| Artifacts (beta) | Cloudflare | Git repositories of files | Yes: commits, branches and forks | Not documented | No | Any Git client | Workers binding, REST, Git |
| Pause, resume and snapshots | E2B | The whole sandbox, files and memory | Snapshots only | No | No | Not documented | SDK |
| Volumes | Modal | Files, seen by others after a commit and reload | No | No: the last write wins | No | Many functions and sandboxes | SDK, a mount |
| Volumes | Daytona | Files on storage backed by S3 | No | No: the last write wins | No | Many sandboxes | SDK, a mount |
| Sandbox persistence and Drives | Vercel | The sandbox's file system | Snapshots only | Not documented | No | Agents in the same Linux group | SDK, CLI |
| AgentFS | Turso | Files, state and history in one SQLite file | Snapshots and an audit log | Not documented | No | Whoever has the file | CLI, SDK |
| Notion MCP | Notion | Pages and databases | Not documented | Not documented | Not documented | People in the workspace | MCP |
| Supabase MCP | Supabase | Postgres tables | Database branching, experimental | Not documented | Yes, the Postgres schema | Anyone with database access | MCP |
| Contents API and GitHub MCP | GitHub | Files in a repository | Yes, Git history | Yes: an update sends the replaced file's sha | No | Anyone with access to the repository | REST, MCP |
| Workbooks | hedwigai | Tables and documents as markdown tabs, or an Excel, Word or Drive file | Yes, every write; any version can be restored | Yes: a write from an older version is refused | Yes, compiled against the workbook's schema | Other agents, and people in the web app | MCP, web app |
Tencent's Agent Bucket: one Space per agent
Tencent Cloud launched Agent Bucket in July 2026. It is a bucket type inside its object storage, COS. A bucket is split into Spaces. Each Space belongs to one agent, one user or one team, and has its own quota, permissions and access token. Agents save reports, images, code and intermediate results there, next to the files users upload. Tencent says its own agent products run on it.
Agent Bucket solves who owns a file and who may read it. Its docs do not describe versions of a single file, a check on what is written, or a refusal of a write from an old copy. A file is bytes at a path.
Few stores keep old versions
Only 5 of the 23 keep earlier versions of a single thing the agent saved: Google’s ADK artifacts, Anthropic’s memory stores, LangGraph’s checkpoints, Cloudflare’s Artifacts and GitHub. The sandboxes keep snapshots of a whole disk at most. Anthropic’s Files API keeps no versions because a file can never change. A changed report is a new file, and a deleted one cannot be recovered.
Fewer refuse a stale write
Two agents, or one agent in two sessions, often work from copies. When the older copy is written back, it erases the newer one. Only 2 of the 23 stop this. Anthropic’s memory stores take a hash of the content the write was based on, and GitHub’s contents API takes the sha of the file being replaced. Modal and Daytona say in their docs that the last write wins.
Almost none check what is written
A file store accepts any bytes. Only 3 of the 23 can check a write against a shape: CrewAI through Pydantic types on a flow’s state, Cloudflare through a hook that can reject a change, and Supabase through the Postgres schema. None of them keeps versions, refuses stale writes and checks each write together.
Where hedwigai workbooks sit
hedwigai is a system of record for agents, reached over MCP. An agent keeps its work in workbooks. A workbook is a set of tabs, each a table or a page of notes, or an Excel or Word file. It can also be linked to a file in Google Drive or Dropbox, and that file stays the record.
Every write makes a new version, and any version can be read or restored. A write says which version it was based on, and it is refused if someone wrote a newer one since. Each write is compiled against the workbook’s schema: its columns, types, required cells, keys and references. A bad cell comes back as an error, the way a compiler answers, and the agent fixes it and writes again. Claude, ChatGPT, Codex and Cursor read the same workbook, and so do people in the web app.
claude mcp add --transport http --scope user hedwigai https://mcp.hedwigai.com
In Claude, ChatGPT or Cursor, add https://mcp.hedwigai.com as a custom connector. The server is in the official MCP registry as com.hedwigai/hedwigai. Sign-in is OAuth, so there is no token to paste.
How we read them
We read each maker’s own documentation, linked from the store’s name in the table. A cell says “Yes” only where the page says so. We left out stores whose only description was a press report or a code sample: Vertex AI Agent Engine’s code-execution files, ByteDance’s Volcengine AgentKit, Alibaba’s AgentRun, Archil and Mesa. Memory services such as Mem0, Zep and Letta are left out on purpose, because they store what an agent remembers rather than what it makes.
More head-to-head evals
| Measure | Particle-1 | DeepSeek V4.1 Flash | GLM-5.3-Flash |
|---|---|---|---|
| Rows right, blind | 91.2% | 89.4% | 90.8% |
| Median seconds per run | 375 | 315 | 144 |
| Measure | GLM-5.3 | DeepSeek V4.1 Flash |
|---|---|---|
| Right, of 75 | 75 | 71 |
| Seconds a question, median | 3.9 | 1.8 |
| Measure | First run | Best since |
|---|---|---|
| Landscape, seconds | 937 | 233 |
| Memo, seconds | 759 | 602 |
| Measure | Qwen3 235B | GLM-5.2 | Gemini 3.8 Flash | Claude Opus 5 |
|---|---|---|---|---|
| Seconds to the plan, median | 27 | 48 | 57 | 94 |
| Plans written, of three runs | 2 | 3 | 3 | 3 |
| Measure | SubQ | jev | Decider | Nimble |
|---|---|---|---|---|
| Found, mean of five columns at section length | 64% | 65% | 66% | 68% |
| Mechanism of action, found at section length | 57% | 51% | 61% | 54% |
| Measure | ask | Apollo | treg |
|---|---|---|---|
| Usable leads | 13 of 40 | 7 of 40 | 6 of 28 |
| Cost per usable lead | $0.037 | $0.27 | $0.23 |
| Measure | ask with jev | ask with Decider |
|---|---|---|
| Questions right, of 92 | 89 | 89 |
| Wrong answers | 1 | 2 |
| Measure | jev | ask |
|---|---|---|
| Our 370 clinical decisions, right | 94.6% | 93.0% |
| PubMedQA, right (best EdgeEvals scored: 62.2%) | 77.0% | 74.4% |
| Measure | ask | Claude Opus 5.5 |
|---|---|---|
| Questions right, of 95 | 59 | 63 |
| Small companies right, of 20 | 13 | 9 |
| Measure | realestate-us | Best of four assistants |
|---|---|---|
| Floor-plan rents right, of 26 | 26 | 23 |
| Amenities claimed the listing doesn't state | 12 | 129 |
| Measure | ask | jev |
|---|---|---|
| Right, of the questions answered | 98% | 79% |
| Wrong answers, of 92 | 2 | 19 |
| Measure | Trained localiser | Best of four frontier models |
|---|---|---|
| Mean overlap with the defect (IoU, 0 to 1) | 0.421 | 0.109 |