A text-based work management system for technologists.

Modern businesses are natively digital, but lack a unified view. Your data is scattered across SaaS tools you don't control, so you piece together answers by jumping between platforms.

Your business is a graph: customers link to projects, projects link to tasks, people link to organizations. Firm lets you define these relationships in plain text files (you own!).

Version controlled, locally stored and structured as code with the Firm DSL. This structured representation of your work, business-as-code, makes your business readable to yourself and to the robots that help you run it.

• Everything in one place: Organizations, contacts, projects, and how they relate.
• Own your data: Plain text files and tooling that runs on your machine.
• Open data model: Tailor to your business with custom schemas.
• Automate anything: Search, report, integrate, whatever. It's just code.
• AI-ready: LLMs can read, write, and query your business structure.

Firm operates on a "workspace": a directory containing all your .firm DSL files. The Firm CLI processes every file in this workspace to build a unified, queryable graph of your business.

The first step is to add an entity to your workspace. You can do this either by using the CLI or by writing the DSL yourself.

Use firm add to interactively generate new entities. Out of the box, Firm supports a set of pre-built entity schemas for org mapping, customer relations and work management. The CLI will prompt you for the necessary info and generate corresponding DSL.

$ firm add

Adding new entity

> Type: organization
> ID: megacorp
> Name: Megacorp Ltd.
> Email: [email protected]
> Urls: ["corp.com"]

Writing generated DSL to file my_workspace/generated/organization.firm

Alternatively, you can create a .firm file and write the DSL yourself.

organization megacorp {
  name = "Megacorp Ltd."
  email = "[email protected]"
  urls = ["corp.com"]
}

Both of these methods achieve the same result: a new entity defined in your Firm workspace.

Once you have entities in your workspace, you can query them using the CLI.

Use firm list to see all entities of a specific type.

$ firm list task

Found 7 entities with type 'task'

ID: task.design_homepage
Name: Design new homepage
Is completed: false
Assignee ref: person.jane_doe

...

To view the full details of a single entity, use firm get followed by the entity's type and ID.

$ firm get person john_doe

Found 'person' entity with ID 'john_doe'

ID: person.john_doe
Name: John Doe
Email: [email protected]

The power of Firm lies in its ability to travel a graph of your business. Use firm related to explore connections to/from any entity.

$ firm related contact john_doe

Found 1 relationships for 'contact' entity with ID 'john_doe'

ID: interaction.megacorp_intro
Type: Call
Subject: Initial discussion about Project X
Interaction date: 2025-09-30 09:45:00 +02:00
Initiator ref: person.jane_smith
Primary contact ref: contact.john_doe

The Firm CLI is available to download via Github Releases. Install scripts are provided to make the process easy.

curl -fsSL https://raw.githubusercontent.com/42futures/firm/main/install.sh | sudo bash

If you don't feel confident running it with sudo, you can:

1. Download the release

   • Go to Github Releases
   • Download the appropriate archive for your operating system and architecture. You can run uname -m in your terminal if you're not sure which one to pick.

2. Extract the archive

tar -xzf firm-[OS]-[ARCH].tar.gz

3. Navigate to the extracted directory

cd firm-[OS]-[ARCH]

4. Run the application

Option A: Run from current directory

./firm

Option B: Install globally (recommended)

# Make executable (if needed)
chmod +x firm

# Move to system PATH
sudo mv firm /usr/local/bin/

# Now you can run firm from anywhere
firm

irm https://raw.githubusercontent.com/42futures/firm/main/install.ps1 | iex

Beyond the CLI, you can integrate Firm's core logic directly into your own software using the firm_core and firm_lang Rust packages. This allows you to build more powerful automations and integrations on top of Firm.

First, add the Firm crates to your Cargo.toml:

[dependencies]
firm_core = { git = "https://github.com/42futures/firm.git" }
firm_lang = { git = "https://github.com/42futures/firm.git" }

You can then load a workspace, build the entity graph, and query it programmatically:

use firm_lang::workspace::Workspace;
use firm_core::EntityGraph;

// Load workspace from a directory
let mut workspace = Workspace::new();
workspace.load_directory("./my_workspace")?;
let build = workspace.build()?;

// Build the graph from the workspace entities
let mut graph = EntityGraph::new();
graph.add_entities(build.entities)?;
graph.build();

// Query the graph for a specific entity
let lead = graph.get_entity(&EntityId::new("lead.ai_validation_project"))?;

// Traverse a relationship to another entity
let contact_ref = lead.get_field(FieldId::new("contact_ref"))?;
let contact = contact_ref.resolve_entity_reference(&graph)?;

This gives you full access to the underlying data structures, providing a foundation for building custom business automations.

Firm is organized as a Rust workspace with three crates:

Core data structures and graph operations.

• Entity data model
• Typed fields with references
• Relationship graph with query capabilities
• Entity schemas and validation

DSL parsing and generation.

• Tree-sitter-based parser for .firm files
• Conversion between DSL and entities
• Workspace support for multi-file projects
• DSL generation from entities

Grammar is defined in tree-sitter-firm.

Command-line interface, making the Firm workspace interactive.

Firm's data model is built on a few key concepts. Each concept is accessible declaratively through the .firm DSL for human-readable definitions, and programmatically through the Rust packages for building your own automations.

Entities are the fundamental business objects in your workspace, like people, organizations, or projects. Each entity has a unique ID, a type, and a collection of fields.

In the DSL, you define an entity with its type and ID, followed by its fields in a block:

person john_doe {
    name = "John Doe"
    email = "[email protected]"
}

In Rust, this corresponds to an Entity struct:

let person = Entity::new(EntityId::new("john_doe"), EntityType::new("person"))
    .with_field(FieldId::new("name"), "John Doe")
    .with_field(FieldId::new("email"), "[email protected]");

Fields are typed key-value pairs attached to an entity. Firm supports a rich set of types:

• String
• Integer
• Float
• Boolean
• Currency
• DateTime
• List of other values
• Reference to other fields or entities
• Path to a local file

In the DSL, the syntax maps directly to these types:

my_task design_homepage {
    title = "Design new homepage"        // String
    priority = 1                         // Integer
    completed = false                    // Boolean
    budget = 5000.00 USD                 // Currency
    due_date = 2024-12-01 at 17:00 UTC   // DateTime
    tags = ["ui", "ux"]                  // List
    assignee = person.jane_doe           // Reference
    deliverable = path"./homepage.zip"   // Path
}

In Rust, these are represented by the FieldValue enum:

let value = FieldValue::Integer(42);

The power of Firm comes from connecting entities. You create relationships using Reference fields.

When Firm processes your workspace, it builds the entity graph representing of all your entities (as nodes) and their relationships (as directed edges). This graph is what allows for traversal and querying.

In the DSL, creating a relationship is as simple as referencing another entity's ID.

contact john_at_acme {
    person_ref = person.john_doe
    organization_ref = organization.acme_corp
}

In Rust, you build the graph by loading entities and calling the .build() method, which resolves all references into queryable links.

let mut graph = EntityGraph::new();
graph.add_entities(workspace.build()?.entities)?;
graph.build(); // Builds relationships from references

// Now you can traverse the graph
let contact = graph.get_entity(&EntityId::new("contact.john_at_acme"))?;
let person_ref = contact.get_field(FieldId::new("person_ref"))?;
let person = person_ref.resolve_entity_reference(&graph)?;

Schemas allow you to define and enforce a structure for your entities, ensuring data consistency. You can specify which fields are required or optional and what their types should be.

In the DSL, you can define a schema that other entities can adhere to:

schema custom_project {
    field {
        name = "title"
        type = "string"
        required = true
    }
    field {
        name = "budget"
        type = "currency"
        required = false
    }
}

custom_project my_project {
    title  = "My custom project"
    budget = 42000 EUR
}

In Rust, you can define schemas programmatically to validate entities.

let schema = EntitySchema::new(EntityType::new("project"))
    .with_required_field(FieldId::new("title"), FieldType::String)
    .with_optional_field(FieldId::new("budget"), FieldType::Currency);

schema.validate(&some_project_entity)?;

Firm includes schemas for a range of built-in entities like Person, Organization, and Industry.

Firm's entity taxonomy is built on the REA model (Resources, Events, Agents) with inspiration from Schema.org, designed for flexible composition and efficient queries.

Every entity maps to a Resource (thing with value), an Event (thing that happens), or an Agent (thing that acts).

We separate objective reality from business relationships:

• Fundamental entities represent things that exist independently (Person, Organization, Document)
• Contextual entities represent your business relationships and processes (Contact, Lead, Project)

Entities reference each other rather than extending. One Person can be referenced by multiple Contact, Employee, and Partner entities simultaneously.

When the entity graph is built, all Reference values automatically create directed edges between entities. This enables traversal queries like "find all Tasks for Opportunities whose Contacts work at Organization X" without complex joins.