With the ability to create projects on Literal AI, you can compartmentalize different parts of your LLM Product.

Creating & inviting members to a Project

To start a new project, navigate to the bottom part and create a new project.

Once a project is created, you can invite others to collaborate:

  1. Go to the project settings.
  2. Invite members by entering their email addresses.
  3. Assign roles to each member.
Invite a collaborator

Invite a collaborator

Role Based Access Control (RBAC)

Project administrators hold the responsibility of assigning and managing permissions for all team members. Permissions in Literal AI are packaged as the following set of roles :

####Supported roles

  • Admin: Full access to the project, including team management, project settings, and all resources.
  • AI Engineer: Manages observability data, API keys, and AI prompts. Cannot alter project settings or team roles.
  • Domain Expert: Reads observability data, scores, and tags data. Limited access to project resources and no API key or project settings management.

Feel free to reach out to us to create a custom role: contact@literalai.com

Permissions per role

AdminAI EngineerDomain Expert
Annotations (scores, tags…)Read & WriteRead & WriteRead & Write
Observability (threads, steps…)Read & WriteRead & WriteRead
Datasets (items, experiments…)Read & WriteRead & WriteRead
Prompt (playground, versions…)Read & WriteRead & Write
API Keys & CredentialsRead & WriteRead & Write
Project & TeamRead & WriteRead

Managing Literal AI API Keys

You can then navigate to the “Settings > General” tab in your project, access the “Literal AI API Keys” section, generate a new API key, and use it in your application.

Managing LLM API Keys

You can find more information on managing your credentials here.

Project-level settings

Project-level settings in Literal AI:

  • LLM Providers: Set up credentials for LLM providers like Groq, Anthropic, OpenAI, and Mistral.
  • Tags: Organize project threads and steps with tags. Add or remove as needed.
  • Score Templates: Craft score templates to assess LLM application performance.

See documentation for setup details or contact support for help.

Project-level settings

Project-level settings