Introspection

Let your engineers code only what your business does differently

Let your engineers code only what your business does differently

Let your engineers code only what your business does differently

Forest reads your tables and their relations. Your engineers start from a map of your data, and your AI agents get the same map through the MCP server.

TL;DR

Forest reads the tables of each database you connect and the relations inside it.

Your engineers skip the data layer and code only the actions your business runs, with every table and field autocompleted.

Your AI agents get the same map through the MCP server, so they know where your data lives before they act.

Your engineers start from a map of your data

Your engineers start from a map of your data

The Forest back-end reads the schema of PostgreSQL, MongoDB or Snowflake at startup, inside your infrastructure. List and detail screens, search, filters and relations are in place before your engineers write a line.

Your tables, live on day one

Every table becomes a collection, with its fields and types mapped. Nothing to declare by hand before your team sees its records.

One click from a customer to their cases

Foreign keys become navigable relationships. From a customer, your team opens their accounts, transactions and cases without a single join.

Search, filter and edit without code

List, search, filter, sort, paginate, create and edit. The operations every backoffice needs work on every collection from the start.

One dataset, a view per team

Forest generates list and detail views for each collection. Each team then shapes its own fields and layout on the same records.

Engineering time goes to your actions

Introspection gives you the records. Forest Actions give you the operations your business runs, coded or vibecoded by your engineers on top.

Schema changes land as a pull request

When your data model evolves, Introspection picks up the change at the next restart. The schema snapshot lives in your repository, reviewed like any other code.

Built for engineering teams

Your data model is the spec. Forest reads it.

Your data model is the spec. Forest reads it.

Engineers code on typed tables

A new action starts from customers, accounts and cases already typed and linked, so the time goes into the business rule itself, a refund threshold or a freeze condition.

A map for your AI agents

An AI agent can only act on data it knows how to reach. Through the MCP server, it gets every collection Introspection read, every relation inside each database and the links your engineers declared between databases.

Structure, not data

Introspection reads table definitions and metadata. Your records stay in your databases, read in place by the Forest back-end inside your infrastructure.

Connect what you run

The databases and warehouses Introspection reads

The databases and warehouses Introspection reads

Frequently asked questions

You still have question ?
Book a conversation with a forest expert

What does Introspection build?

When the Forest back-end starts, it reads the schema of each database you connect: tables, fields, types and the relationships the database declares. Each table becomes a collection in Forest, with standard record operations and default list and detail views.

Does it detect relationships between two databases?

A foreign key cannot point to another database, so Introspection has nothing to read there. Your engineers declare those links in the Forest back-end, one line per relationship. They then work like any introspected relationship.

How does Introspection help us code actions?

Forest Actions are written against the collections Introspection builds. On Node.js, the Forest back-end also generates a typing file, so collection, field and relationship names autocomplete in your editor and in your AI coding agent.

How do AI agents use Introspection?

Before an AI agent acts, it reads through the MCP server which collections exist and how they relate, inside each database and across the links your engineers declared. It finds a customer's accounts, cards and cases on its own.

What happens when our schema changes?

Restart the Forest back-end after a migration and the new tables and fields appear in Forest. The schema snapshot, .forestadmin-schema.json, lives in your repository, so the change shows up in a diff before it reaches production.

Does Forest see our data?

The Forest back-end runs in your infrastructure and reads your databases in place. For SQL databases and Snowflake, Introspection reads catalog metadata only. MongoDB has no declared schema, so the Forest back-end samples a set number of documents to infer one, still inside your infrastructure.

Which databases does Introspection work with?

PostgreSQL, MySQL, MariaDB and SQL Server through the SQL datasource, along with MongoDB, Snowflake and ORM-based datasources such as Sequelize, Mongoose and ActiveRecord. See the full list in the docs.

{"@context":"https://schema.org","@type":"FAQPage","mainEntity":[{"@type":"Question","name":"What does Introspection build?","acceptedAnswer":{"@type":"Answer","text":"When the Forest back-end starts, it reads the schema of each database you connect: tables, fields, types and the relationships the database declares. Each table becomes a collection in Forest, with standard record operations and default list and detail views."}},{"@type":"Question","name":"Does it detect relationships between two databases?","acceptedAnswer":{"@type":"Answer","text":"A foreign key cannot point to another database, so Introspection has nothing to read there. Your engineers declare those links in the Forest back-end, one line per relationship. They then work like any introspected relationship."}},{"@type":"Question","name":"How does Introspection help us code actions?","acceptedAnswer":{"@type":"Answer","text":"Forest Actions are written against the collections Introspection builds. On Node.js, the Forest back-end also generates a typing file, so collection, field and relationship names autocomplete in your editor and in your AI coding agent."}},{"@type":"Question","name":"How do AI agents use Introspection?","acceptedAnswer":{"@type":"Answer","text":"Before an AI agent acts, it reads through the MCP server which collections exist and how they relate, inside each database and across the links your engineers declared. It finds a customer's accounts, cards and cases on its own."}},{"@type":"Question","name":"What happens when our schema changes?","acceptedAnswer":{"@type":"Answer","text":"Restart the Forest back-end after a migration and the new tables and fields appear in Forest. The schema snapshot, .forestadmin-schema.json, lives in your repository, so the change shows up in a diff before it reaches production."}},{"@type":"Question","name":"Does Forest see our data?","acceptedAnswer":{"@type":"Answer","text":"The Forest back-end runs in your infrastructure and reads your databases in place. For SQL databases and Snowflake, Introspection reads catalog metadata only. MongoDB has no declared schema, so the Forest back-end samples a set number of documents to infer one, still inside your infrastructure."}},{"@type":"Question","name":"Which databases does Introspection work with?","acceptedAnswer":{"@type":"Answer","text":"PostgreSQL, MySQL, MariaDB and SQL Server through the SQL datasource, along with MongoDB, Snowflake and ORM-based datasources such as Sequelize, Mongoose and ActiveRecord. See the full list in the docs."}}]}

Introspection

LEVEL UP YOUR OPS GAME

One control plane. Every action traced for human, agent, BPO, LLM, workflow.

Introspection

LEVEL UP YOUR OPS GAME

One control plane. Every action traced for human, agent, BPO, LLM, workflow.

Introspection

LEVEL UP YOUR OPS GAME

One control plane. Every action traced for human, agent, BPO, LLM, workflow.

The operational infrastructure regulated companies grow on

Copyright ยฉ 2026 Forest

Design by Alasta & Built by Reiya Studio