DIN is the system of record for AI work, for any agent. It captures what an AI agent did at the tool-call boundary, grades each claim as evidence-supported or asserted, and keeps that record on your infrastructure. The free build is one field, up to 3 people and 1 AI agent, no signup, no card; the self-hosted plans from Starter to Small Business are on sale, monthly.
A log tells you the agent ran. It does not tell you whether its claims survived contact with the evidence. A transcript of a claim is still a claim, written in the same confident voice whether the work held or not. DIN keeps the grade: what was proved, what was only stated, and what was later contradicted.
The record lives in a field: one canonical store per tenant, where records are placed by what they rest on rather than filed by date. It runs on a machine you control, and the field listens on your own machine only: agents on the same machine connect through the plugin or the CLI, cloud chat history comes in through import, and hosted access is set up with you.
Five steps, the same for every producer, whether the work came from a Claude Code session or a Codex rollout:
Two questions, asked separately and never merged into one badge: how a record was obtained, and what it turned out to be worth. How it was captured is one of receipted, logged, inferred or mixed. What it proved is one of proved, asserted or refuted. A refuted claim stays in the record, marked, rather than being quietly removed.
An inferred record never renders as a receipted one, anywhere in the console, at any zoom. Where the record cannot settle a position on its own, a person rules, with a name and a date on the stamp.
One meter: fields. A field is a store, a team, a project, a client: one container of records, opened by one token. You are not billed for volume, for seats, for storage or for inference: grading runs on inference you already have (your assistant CLI, your own API key or a local model), so there is no usage line from us.
From DIN 1.0.39 the field count is enforced: creating a field past your plan's number is refused, with the plan, the number and the way up named, and fields you already have keep working. Client orgs (Agency: 5) and ops layers are enforced the same way, from the signed licence; people on each plan are shown, not enforced.
The free build is one field, up to 3 people and 1 AI agent, with no signup and no card, and it is available to download today for Linux x86-64 (verified). macOS is available on request. History import (graded up to the Reader level, in a 40,000-character window), recall, the scoped brief, the local console and the dependency inventory are free. Live capture, every reading level and the whole-chat read, and the dependency watch's delivery are paid.
Paid plans are priced by number of fields, in USD per month. Self-hosted: Starter $29 (6 fields), Freelancer $99 (15), Team $199 (30), Small Business $299 (90), Agency $399 (90 fields across up to 5 client orgs). Enterprise (180 fields and up) is a conversation, not a checkout. Hosted, where we run it: from $50 (Starter) to $599 (Agency). The self-hosted plans from Starter to Small Business are on sale, monthly, and the pricing page opens a checkout on Stripe's own payment page. Agency and hosted are set up with you: contact us from the pricing page.
DIN grades claims against the evidence that produced them, so the evidence must still exist. Nine sources import today: Claude Code, Codex, Gemini CLI, Qwen Code, the Cline family, claude-mem, Grok, ChatGPT (its data export), and Zep / Graphiti (through din-import-graphiti). A tenth, claude.ai with the Claude desktop chats, imports from the claude.ai data export (din import claude-ai): each conversation is kept, secrets scrubbed, and placed ungraded at no cost, and --grade grades it. Importing is not grading: grading runs on your own inference, on any of three paths: your assistant CLI (Claude Code or Codex), your own API key (Anthropic, OpenAI, or any OpenAI-compatible endpoint) or a local model (Ollama, vLLM, llama.cpp). Claude Code is the path DIN's grading was measured on; the Codex, key and local paths are built and tested on stand-ins, and none of them has been validated against a real vendor yet. On Codex or an OpenAI key, gpt-6-sol at medium effort reads at Reader on Free (din grade on --effort sets the effort; Analyst and up are paid). Grading on the Google Gemini CLI is turned off, because DIN cannot lock its tools down. The matrix at /compatibility says what works with each agent and source. Other tools that keep the original conversation (Letta, LangGraph, OpenAI Sessions, AWS AgentCore events, Azure Foundry conversations, Google Agent Engine, Supermemory, Khoj, Basic Memory, SpecStory) can be graded once a mapper for them ships. Tools that store only extracted facts cannot be graded, because the thing a grade would point at is gone.
Live capture, a paid feature: the installer asks which agent to capture, Claude Code, Cursor, OpenCode or Codex (or takes --agent). Gemini CLI, Qwen, Cline and Grok history import. Recall into the agent reaches Claude Code, Cursor, OpenCode, Codex and Gemini CLI once wired; anywhere else it is din recall in a terminal, and any agent runtime can read the field over its HTTP API with the token, and a person can stamp a record.
The free build is: one field, up to 3 people and 1 AI agent, with graded history import (up to the Reader level, in a 40,000-character window), recall, the scoped brief and the local dashboard, no signup and no card, running entirely on your own machine. Paid is the same download plus a key that turns on live capture, every reading level and the whole-chat read, several agents on one field, more fields, the dependency watch and the paid panels. What is on sale today is on the pricing page.
On infrastructure you control: your laptop or your own box. The installer starts one server on loopback behind a token it mints on your machine, and the field listens on your own machine only: agents on it connect through the plugin or the CLI, and cloud chat history comes in through import. Grading runs on your own inference: your assistant CLI (Claude Code or Codex), your own API key (Anthropic, OpenAI, or any OpenAI-compatible endpoint) or a local model (Ollama, vLLM, llama.cpp); DIN never pays for or sees those calls. A free install never checks in. A paid key is verified when you run din activate, and the signed entitlement it fetches belongs to that install and lasts 14 days; a paid install checks in, with a random install id, when you run din activate --renew and at most once a day when its field or console starts (or with din autostart).
Linux x86-64, verified. macOS is available on request (write to support@dinctrl.com): it is not on the download page, and the Linux installer says so on a Mac rather than failing halfway through a copy.
Not yet. The Community Edition source is coming: the plumbing will be licensed Apache-2.0 when it is published; binaries today. Where its public repository is published, the Get DIN page links to it, and until then it shows no link. The compiled grading engine and the commercial services are proprietary.
No. The console is real and running, and every count, trust mark and flag is read live from a field. The records in that field are invented: Brindlewood Trading Co. is a fictional company, nothing here is anyone's real data, and the console is read-only.
Download the free build · Pricing · Docs · For AI agents
Updated 2026-11-16.
Real: this console, the field behind it, and every number on every screen. Nothing on this page is a mock-up, a screenshot or a canned response. Each panel is a live read of a running FieldDIN field, and a panel whose endpoint does not exist says so instead of drawing something plausible.
Not real: the records in that field. Brindlewood Trading Co. is a fictional company, and its eight divisions, from Storefront to Compass Data Quality, are invented. The four people who appear as authors and reviewers are invented, and so is every path, command, test output and refutation. The data set was generated from a scenario, and every evidence address in it is checked against its own transcript before it is loaded.
Read-only. This console proxies GET requests to a fixed list of read routes and holds no path to the field's one write door. Nothing you click can change a record, and nothing you type leaves the page.
This is FieldDIN, running, reading a live field over the network and grading what it finds. The grading is the product, so none of it has been softened for a demo. What has been made up is the work being graded.
DIN keeps a graded record of what your AI agents did: every claim beside its evidence.
DIN connects each recorded claim to its evidence, keeps contradictions, and shows what later work depends on it. Explore a running console using synthetic records. DIN runs on Linux; macOS on request.
Keep your memory; DIN adds evidence grades, contradictions and the why. It reads what your tool already holds and grades it against the conversation that produced it.
din-import-graphiti)What DIN can read: the compatibility page Download the free build
| Receipted | regime | Produced by a governed run (SpinDIN). Every step carries a keyed receipt (HMAC). The strongest evidence DIN has. |
|---|---|---|
| Logged | regime | Captured live with the tool calls and their results present. What the agent ran and what came back is on record. |
| Mixed | regime | A record whose claims came from more than one kind of capture. |
| Inferred | regime | Reconstructed from a conversation history or an export, without the tool results (web chat exports, for example). Useful, but read with care. |
| Tested | claim state | Supported by something that was executed in that session. |
| Observed | claim state | Read directly from a file or tool output, nothing executed. |
| Externally confirmed | claim state | Supported by an outside source that was actually consulted. |
| Argued | claim state | Reasoned, but not shown by a result. |
| Asserted | claim state | Stated, with nothing behind it. |
| Disputed | claim state | Contradicted or contested with substance. |
| Field | where records live | Your body of records for one project, team or company. It lives in a folder on your machine and is served by a small local server. |