Work you can check.
A system of record for the spreadsheets and documents your AI edits, checked before every save.
First workbook free. Reading and writing workbooks costs nothing.
Listed in the Claude Connectors Directory · MCP Registry · mcp.so, verified
Which are you?
For Claude Code, Codex, Cursor and any MCP agent
Agent memory that outlives the session.
- CLAUDE.md and memory files stay on one machine. Workbooks hold decision logs, task status, handoffs and results that any session, agent or teammate on your account can read.
- Every write is a new version; a write made on an old version is refused, never merged.
- Each write is checked against the workbook's schema, the way a compiler checks code, with an error for each bad cell.
- An MCP server, connected in one line: claude mcp add --transport http hedwigai https://mcp.hedwigai.com
For people and teams
Let Claude or ChatGPT edit your Excel and Word files in place.
- Your files stay in Google Drive or Dropbox, and each edit is written back there. No downloading and uploading again.
- Every edit is checked against the file's schema before it is saved; an edit that breaks it is refused, with the cell that broke it.
- Every version is kept; an edit made on an old version is refused, never merged.
- Set up in the AI app you already use. The first workbook is free.
Try it now, nothing to install.
Who imports more?
Two US importers of the same product: guess which filed more shipments, from customs records.
Try it →Will your business plan work?
Every order your processes can run in, checked by TLC: where it breaks, the fix, and what it earns.
Try it →Where will AI models be in a year?
The best score any model has reached, forecast up to two years out, and what a model at that score can do.
Try it →Push a neuron in Qwen3.5-4B
Pick one of 13 personalities, ask a question, and see the plain answer next to the pushed one.
Try it →Two characters, one neuron each
Give each a personality, set a scene, and watch them talk and move. Change one while it runs.
Try it →Solo agent against agent teams
Three setups tune the same kernel with 40 calls each, replayed call by call.
Try it →Where does an AI agent quit?
Each moment Qwen3.5-9B could stop in a live run, resampled 16 times: how often it quit, and what it wrote.
Try it →Compile an agent's rules into its model
Four rules in TLA+, checked by TLC and compiled into edits of Qwen3.5-0.8B, which runs in your browser.
Try it →Train a small model on your examples
Give prompts and labels; Qwen3.5-0.8B gets new neurons fitted by least squares in about 10 seconds.
Try it →Does the weight edit stop the loop?
One math problem answered by Qwen3.5-0.8B as released and as an AI agent edited it, side by side.
Try it →Can a monitor catch a pushed neuron?
A monitor reading Nanbeige4.3-33B's thinking ranks a pushed run above a plain one in 91 pairs of 100.
Try it →Does one neuron catch reward hacking?
The claim tested with controls, and the detector run on an answer of your own.
Try it →
01 · MCP for knowledge work
A compiler for knowledge work.
Your spreadsheets and documents stay in your Google Drive or Dropbox; the file is the record. Every AI assistant, and every run it starts, reads and edits that file through one layer, the workbook. Its schema declares each table's columns, types, required cells, keys and references, so the tooling that keeps code correct now keeps your operating records correct.
- Accurate
- Every edit is checked against the schema before it is saved. A change an assistant proposes is checked on a copy first and redone with the errors, so you are only asked to approve one that passes.
- Resilient
- Every version is kept and can be restored. An edit based on an old version, or over a file someone changed in Excel meanwhile, is rejected, never merged over.
- Fast and cheap
- An assistant reads and edits one table or section by its address, not the whole file, so a change costs only what it touches.
Works inside Claude, ChatGPT, Codex and Cursor.
Listed in
A short answer from your workbooks and programs, with the source it rests on. Nothing to wait for.
Is Moderna a biotech with a public float above $5B?
Your agent
ask · Is Moderna NASDAQ:MRNA a biotech company with public float above $5B?
{
"answer": "yes",
"claims": [{ "figure": "public float", "value": "$9.6B",
"as_of": "2026-06-28" }],
"sources": ["MRNA 10-K, cover page",
"biotech-diligence-7fq2 · §Moderna"]
}Yes. Moderna's public float was $9.6B as of June 28, per the cover of its 10-K.
02 · Interpretability experiments
Why a model behaves as it does, measured inside it.
Interpretability explains a model's behaviour from its internals: which neurons and features respond to an input, and how the output changes as the input, the prompt or the environment changes. We test published claims on open-weight models with causal methods (ablation, activation steering, probes and weight edits), each against a control, and publish every number, including the claims that failed. You can rerun each one.
59 of 100
Does one neuron catch reward hacking?
The neuron posted as a reward-hacking detector in Qwen3.5-4B tells a hack from an honest answer in 59 pairs of 100. Another neuron, found with a held-out search, does it in 99.
Run it yourself →9.2% vs 8.6%
Does emergent misalignment run through one neuron?
Switching off the neuron that makes the model hostile leaves 9.2% of answers misaligned. A random control neuron gives 8.6%. The claim does not hold.
Run it yourself →91 of 100
Can a monitor reading the thinking catch a pushed neuron?
A Qwen3.5-9B monitor reading Nanbeige4.3-33B's chain of thought catches a full personality push in 91 pairs of 100, and a half push in 65.
Run it yourself →22 vs 7
Does a weight edit stop a model looping?
On prompts with no math, 22 of 32 answers from the released Qwen3.5-0.8B ran to the 8,192-token limit. The edited copy did it 7 times.
Run it yourself →Few example workbooks
VK2735 competitive landscape: the 8 closest obesity programs
Tirzepatide, CagriSema, retatrutide, orforglipron and four more against Viking's VK2735: phase, latest weight-loss readout, next catalyst and deals, every row cited.
29 sources · 25 minutes
Read the workbook →US importers of bulk black tea from India, from customs records
Four US buyers with the bills of lading that show it, who they buy from, what gates a new supplier, and the leads not worth chasing yet.
10 sources · 8 minutes
Read the workbook →Early-stage AI startups that filed a Form D in the last 30 days
Seven US AI companies raising in September 2026: the instrument each filing names, amount sold, investors and the people on it, each cited to the Form D on EDGAR.
17 sources · 13 minutes
Read the workbook →Tested on real work. Every score published.
Does Physical Agent aim its tool calls at the object?
Physical Agent writes a robot tool call on 400 of 400 frames, but moving the object 16 cm leaves its targets where they were.
| Measure | Physical Agent | Qwen3.5-4B |
|---|---|---|
| LIBERO frames answered with a tool call | 400 / 400 | 18 / 400 |
| Target movement per cm of object movement (median) | 0.02 cm | — |
Frontier models by scale, 2021–2026: every compute band climbs at the same pace
Each band of training compute gains 12 to 20 capability points a year; a lab's smallest model trails the best by about seven months.
| Measure | Value |
|---|---|
| ECI points gained per year, by compute band | 12 to 20 |
| Delay for a band ten times smaller to match | 3 to 13 months |
Particle-1, DeepSeek V4.1 Flash and GLM-5.3-Flash in Claude Code with hedwigai workbooks
Particle-1, DeepSeek V4.1 Flash and GLM-5.3-Flash, 180 runs graded blind, rows right: 91.2%, 89.4% and 90.8%, intervals overlapping.
| 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 |
Why agents quit early
Qwen3.5-9B never stops at an ordinary turn of an agent run; immediately after context compaction it stops about half the time.
| Measure | Qwen3.5-9B | DeepSeek V4.1 Flash | Particle-1 | GLM-5.3-Flash |
|---|---|---|---|---|
| Runs that quit early | 3 of 4 | 3 of 31 | 3 of 29 | 0 of 33 |
| Calls that move to another idea basin | 33% | 63% | 57% | 56% |
Who imports it? Four workbooks that keep the answer, from US customs records
Marin Mountain Bikes filed more bicycle shipments than Giant or Trek; 572 companies imported black tea.
| Measure | US customs records |
|---|---|
| Most bicycle shipments | Marin Mountain Bikes, 5,711 |
| US black tea importers, 2026 | 572 |
Where do AI agents save their work? 24 artifact stores compared
Of 23 artifact stores for agents, 5 keep versions, 2 refuse a stale write and 3 check what is written; none does all three.
| Measure | Tencent Agent Bucket | Anthropic memory stores | hedwigai workbooks |
|---|---|---|---|
| Old versions kept | Not documented | Yes, 30 days | Yes, every write |
| Stale write refused | Not documented | Yes, by content hash | Yes, by version |
Questions, answered.
What does hedwigai do?
It is a system of record for the spreadsheets and documents an AI edits. Claude, ChatGPT, Codex, Cursor or any MCP agent keeps them as workbooks, optionally linked to a file in your Google Drive or Dropbox. Every write is a new version, a write made on an old version is refused rather than merged, and every edit is checked against the workbook's schema before it is saved.
Why keep an agent's state in hedwigai rather than a local file or its memory?
A local file stays on one machine and is overwritten by whoever writes last; an agent's memory is private to it. A workbook is found by the next session and by other people and agents, keeps every version, refuses a write based on an old version, and holds the shape its schema declares, so a tracker or a decision log does not drift.
How do I connect my agent?
In Claude Code: claude mcp add --transport http hedwigai https://mcp.hedwigai.com, then /mcp to sign in, or install the hedwigai plugin with /plugin marketplace add un-seen/hedwigai-skills. In Claude, add it from the Connectors Directory. Codex, Cursor, Hermes and any other MCP app add https://mcp.hedwigai.com as a server.
Where do my files live if I connect the MCP?
In your Google Drive or Dropbox, at the same path. The file is the record: a run reads the latest version first and writes each new version back to it. If someone changed the file in the meantime, the edit is refused rather than written over theirs, and every version can be restored.
Which AI apps does it work with?
Claude, where it is listed in the Connectors Directory; ChatGPT, where the plugin is in review; and Claude Code, Codex, Cursor, Hermes or any other app that adds an MCP server. It is also listed in the MCP Registry, on Glama and on mcp.so.
What does it cost?
The first workbook is free. Reading, searching and writing workbooks is free; research runs, started only when you ask for research, are billed to the account by what they use.











