ResourcesFeature Overview

Feature Overview

A product-wide overview of Giga's current core capabilities and how they fit together.

Giga brings company context, reusable AI work, and connected systems into one permission-aware product model.

The core idea is simple: keep the company’s durable context and access rules in Giga, then use that context from the interface and AI model that fits the work.

Company Brain

Company Brain is Giga’s permission-aware company context layer.

It can combine relevant information from:

  • workspace and Group guidance
  • Tracks
  • Context Notes
  • Observations
  • Current Track State
  • files
  • Connections and live external data
  • Agents
  • Skills
  • Workflows
  • user-level context

Giga retrieves context according to the current user, permissions, and task.

Video placeholder: Company Brain

Add the approved Company Brain product video here.

Chats

A Chat is the direct working surface between a user and Giga.

Use a Chat to ask questions, reason through work, create or edit supported work product, use Connections, and take supported actions.

A Chat can link to at most 1 explicitly selected Track.

Track selection should never be inferred silently because it affects where observations can become retrievable.

Groups

Groups are Giga’s main placement and permission scopes.

Groups can contain or scope resources such as:

  • Tracks
  • Agents
  • files
  • credentials and Connections
  • shared context

A user can access shared Groups they belong to plus their own private Group.

Use a Group when there is a real team boundary, shared area, or top-level access scope.

Tracks

A Track is Giga’s core primitive for work that continues over time.

Use a Track when context, decisions, files, and progress should accumulate around one effort.

A Track’s durable context is split across several layers:

LayerPurpose
Context NoteStable background and standing instructions
ObservationsAppend-only developments and learned context
Current Track StateCurated current position and decision record
FilesAuthored work product
Linked ChatsWorking history around the Track

Track visibility currently follows the Group it lives in.

Context Notes

A Context Note is the durable standing background for a Track.

Use it for things such as:

  • what the Track is about
  • stable constraints
  • durable terminology
  • standing instructions
  • handoff context that should continue to matter

Fast-changing progress belongs in Memory or Current Track State instead.

Memory

Track memory has 2 important layers.

Observations preserve meaningful developments over time.

Current Track State keeps the curated view of where the Track stands, how it got there, what was ruled out and why, and what remains open.

These layers serve different jobs and should stay distinct from authored files.

Files and databases

Files are authored work product.

Confirmed file capabilities include Markdown, text and binary Track files, code, webpages, and visual artifacts where the current surface supports them.

For structured growing records, Giga can use SQLite or a hosted Giga database when available.

Specific rich file-format claims should follow the current live capability rather than a fixed universal format list.

Connections

A Connection gives Giga permissioned access to an external system.

Connections can represent:

  • apps
  • APIs
  • databases
  • MCP servers
  • browser sessions
  • messaging systems
  • other supported services

A Connection can provide current information and, where the provider and Giga permissions allow it, support actions.

For custom Connections, Giga currently prefers the strongest supported self-serve route in this order:

  1. MCP
  2. OAuth with supported self-serve registration
  3. API key or PAT
  4. browser login

Credentials stay in the secure connection layer rather than normal Chats, files, or Context Notes.

Agents

An Agent is a named reusable AI role.

An Agent can have:

  • a role
  • instructions
  • durable context
  • access to permitted Connections
  • Skills

Use an Agent when a reusable operating identity should work across multiple tasks or people.

Its actual access stays governed by the live Group and permission model.

Skills

A Skill is reusable instructions and supporting context for a specific kind of task.

Use a Skill when the reusable asset is primarily how to do something.

Examples include a research method, review standard, handoff format, or task-specific operating procedure.

Skills can be owner-only or explicitly shared into Groups.

Workflows

A Workflow is a reusable multi-step process.

The current Workflow model supports:

  • a name
  • an optional description
  • ordered meaningful steps
  • plain-language instructions
  • optional runtime inputs

Current input types can include text, URL, number, choice, and boolean where exposed by the live Workflow surface.

Keep Workflow steps focused on meaningful stages rather than individual tool calls.

Pulses

A Pulse is recurring scheduled work.

Use a Pulse when Giga should run a check or task automatically on a schedule.

Current Pulse scheduling supports:

  • shared named schedule slots
  • custom schedules
  • daily cadence
  • weekly cadence
  • monthly cadence
  • intra-day cadence

A Pulse can target an Agent, target a Track, or run without a target.

Creating recurring work should be confirmed before it is scheduled.

Permissions

Permissions run through the entire product.

Important layers include:

  • workspace membership
  • workspace role
  • Group membership
  • resource ownership
  • Connection sharing
  • external provider permissions
  • current live tool permissions

A resource existing in the workspace does not automatically make it available to every user or Agent.

Learn about Permissions →

Use Giga anywhere

Giga’s official user-facing surfaces are:

Surface presentation can differ. The underlying Giga workspace remains the authority for context and access.

Model flexibility

Model choice and provider flexibility are a first-class Giga product area.

Giga’s workspace context, permissions, Tracks, Skills, Workflows, and Connections are conceptually separate from a single model provider.

Exact providers, model availability, plan access, enterprise agreements, custom endpoints, and bring-your-own-model options should be verified from the current product configuration before they are documented as available.

Video placeholder: Model Flexibility

Add the approved Model Flexibility product video here.

Choosing the right feature

A quick rule of thumb:

You need...Use...
Direct ad hoc workChat
A team or access boundaryGroup
Ongoing work with historyTrack
Stable Track backgroundContext Note
A named reusable AI roleAgent
A reusable task methodSkill
A reusable ordered processWorkflow
Recurring scheduled workPulse
Live external-system accessConnection
Growing structured recordsDatabase
Auth or secretsSecure Connection flow

See the full Choosing the right Giga feature guide →

Image placeholder

Full product architecture diagram showing Company Brain and permissions underneath Groups, Tracks, Chats, Agents, Skills, Workflows, Pulses, Connections, Files, and the supported user-facing surfaces.