Enterprise AI · Private by construction
Topsoil gives engineering and business teams a capable AI assistant without handing your source code, documents or prompts to someone else's platform. Inference runs through the endpoint you configure, inside the account, region and network you already control.
No vendor-operated inference · No prompt store · No telemetry stream
Use the leading model families — OpenAI, Anthropic Claude, xAI Grok, Amazon Nova and Meta Llama among them — and change your mind later. Models are selected from what your own account enables, so a new one becomes available the day you turn it on.
Topsoil runs against your account, your region, your network path and your IAM policy. That is what makes data sovereignty real rather than contractual: the work never transits a platform operated by us.
There is no analytics pipeline, no usage beacon and no prompt collection leaving your environment. Nothing about your code, your documents or your people is reported anywhere.
Most AI tools ask you to trust a policy document. Topsoil removes the question: there is no vendor-operated control plane in the path, so there is nothing to trust us about.
Where it runs
Engineers live in a terminal and an editor. Analysts, marketers and proposal teams live in an application window. Topsoil is the same product and the same governance in all of them.
A full workspace for macOS and Windows: projects, a context library, documents, diagrams, shared channels and a built-in console. The natural home for people who do not live in a terminal.
The whole agent in a terminal — on a laptop, a build box, a bastion host or an air-gapped workstation. Same models, same permission model, no desktop required.
A bridge, not a second agent. It lends the terminal agent your unsaved buffers, selection, open files and diagnostics, then opens a diff of what the turn wrote. The agent keeps running in the integrated terminal.
Inside the workspace
Every answer shows its sources. The side panel lists the working folder, each attached file, how much of the context window is spent and what the turn cost — so context is something you inspect, not something you hope about.
~/Projects/route-planner
A shell and a chat, on the same folder. git log and
git diff --stat in one pane, the conversation about that change in the other — both
inside the desktop app, with no window to switch to.
Software and AI engineering
Topsoil reads the repository, proposes the change, runs the command, opens the diff and verifies the result — with a permission gate in front of anything that touches your machine.
Inspect and edit files, follow language-server intelligence, and search a local semantic
index that respects .gitignore and never uploads your source to build itself.
Run commands and Git operations under allowlists and path protections. Commands run without a shell, exactly as shown to you before they run.
Apply different models to explore, plan, implement, review and verify. Use a fast model to search and a strong one to reason, within a single piece of work.
Take an issue through isolated worktrees to a reviewed pull request. Backlogs can be GitHub Issues, Jira Data Center with Bitbucket, or a CSV export on disk.
Verify runs the changed test with the rest of the work reverted and requires it to fail first. A test that passes without the change does not count as done.
Add MCP servers over stdio, HTTP or WebSocket, enable them per project, and keep every secret host-side where the model never sees it.
Control
Plan mode stays read-only until you approve. Every edit and every command surfaces first — with the diff, the path and the exact command — and you choose once, for that path, for that tool, or for the rest of the session.
− const entry = etaCache.get(route.id);+ const entry = etaCache.get(`${route.id}:${route.revision}`);
$ topsoil models Models in us-east-1 (refreshed now) OpenAI openai.gpt-5.6-terra tools · docs openai.gpt-5.4 tools · docs Anthropic * us.anthropic.claude-sonnet-4-5 tools · docs · images us.anthropic.claude-opus-4-5 tools · docs · images xAI us-gov.xai.grok-4.6 tools · docs · images Amazon us.amazon.nova-pro-v1:0 tools · docs · images * default model
Model flexibility
Topsoil is not a wrapper around one vendor's model — OpenAI, Anthropic, xAI and the rest sit side by side. It lists what your account actually offers, groups it by vendor, and lets a person pick from a scrollable list rather than pasting an identifier. Change the default for a session, a project or a workflow stage.
Privacy, security and data sovereignty
Topsoil was built for organizations that cannot send engineering context or commercial documents to a multi-tenant platform — regulated industries, defense programs and anyone whose data has a jurisdiction.
Prompts, source, documents and output stay inside the account, region and network path you configure. There is no Topsoil-operated service in the request path and no prompt store anywhere outside your environment.
No analytics, no usage beacons, no crash payloads carrying your content. The absence is architectural: there is no pipeline to disable, because one was never built.
Authenticate with IAM roles and keys, AWS profiles, IAM Identity Center (SSO), or a gateway bearer token. Access, logging and retention stay under your existing policy.
Public endpoints, custom HTTPS endpoints, or reachable PrivateLink and VPC paths — including isolated and government cloud partitions where that is the requirement.
Agent capabilities start switched off. Turning one on is a deliberate act with a confirmation, so a workspace is never more capable than someone intended it to be.
Agent tools can read and modify files, run commands, and start local servers. Only enable for projects you trust.
$ /report Route Planner 2.4 readiness write document docs/route-planner-2.4-readiness.pdf
Saved into the working folder. PDF, Word, Excel and Markdown.
Beyond engineering
Contracts, proposals, pricing models and campaign plans are exactly the material an organization cannot paste into a public assistant. Topsoil writes real files — PDF reports, Word memos, Excel workbooks and Markdown — against a private library of your own documents.
The terminal
On a build server, a jump host or a locked-down workstation, a desktop application is not an option. The TUI is the full agent — same models, same permission gates, same project context — in a terminal.
Real work, in a terminal. Search the repository, read the file, explain the defect and propose the change — with the token count and cost on the footer the whole time. See the full terminal walkthrough →
At a glance
| Capability | What it means in practice |
|---|---|
| Private inference | Requests go to the Bedrock, custom HTTPS or gateway endpoint you configure. No vendor-operated inference control plane. |
| Model selection | OpenAI, Anthropic, xAI, Amazon and Meta among them — every family your account enables, grouped by vendor and selectable per session, project or workflow stage. |
| Coding agent | Read and edit files, run approved commands, use Git, language servers, browser, web and document tools. |
| Plan-first governance | Read-only plan mode, tool risk classes, path protections, command allowlists, explicit approvals. |
| Project context | Projects, attachments and a per-identity context library backed by your own storage. |
| Local semantic search | An optional local index of your workspace that honours .gitignore and is built without uploading source. |
| Documents | PDF reports, Word memos, Excel workbooks, Markdown and rendered diagrams, saved into your working folder. |
| Integrations | MCP servers over stdio, HTTP or WebSocket, enabled per project, with secrets kept host-side. |
| Delivery | Backlog to pull request through isolated worktrees, with a verification step that must prove the test fails first. |
| Collaboration | Shared projects and channels running in your account, sharing the stream and selected files — not credentials, terminals or raw tool output. |
| Surfaces | Native desktop for macOS and Windows, a cross-platform terminal client, a VS Code context bridge, and a customer-hosted web option. |
| Deployment | Your cloud account and region, including isolated and government partitions. Nothing is hosted by us. |
Topsoil is powered by Amazon Bedrock and runs on AWS, including GovCloud — in your account, under your controls.
Get started
A proof of concept runs in a secure enclave with your models and your material, and ends with findings you can act on — not a demo environment and a slide.