Legal counsel your coding agent can call.
Legal review at the speed you build. Your agent answers counsel's questions straight from the codebase, with evidence, in minutes. You get final documents and a clear path to launch. Your agent gets a build spec and a test checklist that show it how to ship the feature, not just what could go wrong.
Built for three kinds of products
Most AI products are all three at once. The law reaches each layer differently.
AI features
Models in your product: what goes into prompts, what the provider keeps, what users are told, and which state AI and chatbot laws apply.
Agents
Products that act for users. Every agent acts under a mandate: what the user authorized, how you prove it, who pays when it goes wrong, and what the sites it operates allow.
The app around them
Privacy policies, terms, health and financial data, consent flows, deletion, and app store rules, all written to match what the code does.
How an engagement works
The agent handles everything except the parts that must be human: the attorney's judgment, and your signature on the engagement.
Why evidence beats assumptions
A policy that promises something the product doesn't do is a false statement. The questions that decide what your documents can say are questions about code.
- Can the agent spend money without a fresh confirmation?Answered by quoting the code that enforces the gate, not by recollection.
- What user data reaches the model provider, and is it kept?Traced from the prompt builder to the API call, with the retention setting quoted.
- What happens when a web page tells the agent to do something?The agent shows how untrusted content is handled, and counsel decides what the terms can promise.
- Does anything collect data before consent?The agent traces SDK initialization and tells counsel exactly where.
Traditionally these questions travel founder to developer to lawyer and take weeks. An agent answers them in minutes.
Services
Fixed fees, stated turnaround, nothing paid up front. Your agent gets the exact fee before you engage, and it's confirmed in your engagement letter.
Loading catalog…
What you get back
Every deliverable comes in two forms that say the same thing.
- Final documentsTerms, agent authorization terms, privacy policies, AI disclosures. The legal work product.
- Privileged memoCounsel's advice for your decision-makers: the risks that matter, and the legal reason behind each build item. It stays out of your repository, which protects privilege.
- Build specConfirmation gates, action logs, consent screens, provider settings, deletion flows. Each item says what to build and what it achieves, written to be safe to commit.
- Verification checklistEach row tests one statement in the published documents against the shipping build. A failing row means a document says something false.
- Counsel triggersA file in your repo listing changes that need review first: a new model provider, a new agent action, spending money, teen users. Your agent checks it before it builds.
The sample shown is a fictional travel-booking agent.
Loading…
Guides
Legal information for builders and their agents. Each guide has a Markdown version and a checklist an agent can run against a codebase. All guides
AgentsCan my AI agent buy things for users?
AgentsWhat legal work does an MCP server or agent tool need before launch?
AgentsHow do I keep legal advice privileged when my agent talks to my lawyer?
AgentsWhat should I do when a user says my agent did something they didn't authorize?
AgentsHow should my agent handle users' passwords and logins?
AgentsCan my AI agent give legal, health, or financial help?
AgentsWho is responsible when my agent uses another company's website?
AI featuresDo I have to tell users they're talking to an AI?
AI featuresWho owns AI-generated output, and what should my terms say?
AI featuresCan I train or fine-tune a model on my users' data?
AI featuresWhat should the terms of service for an AI product cover?
AI featuresDoes Colorado's automated decision-making law apply to my product?
AI featuresCan I send user data to an LLM provider?
App privacyCan my iPhone app send user data to a third-party AI model?
App privacyDoes Washington's My Health My Data Act apply to my app?
App privacyIs it risky to use session replay or analytics SDKs in my app?
App privacyWhat changed in the COPPA rule, and does it apply to my app?
For agents
Start with /llms.txt. Connect the MCP server, or call the REST API directly.
/llms.txtWhat we do, rules, how to engage
/api/services.jsonServices, fees, turnaround
/api/intake-schema.jsonWhat counsel needs to know
/api/deliverable-schema.jsonWhat comes back
/api/counsel-triggers.jsonWhen to call before you build
/openapi.yamlREST API
{
"mcpServers": {
"counsel": { "type": "http", "url": "https://mandatelaw.ai/mcp" }
}
}
Ground rules
These are professional obligations, not legal fine print.