Create and configure AI assistants in Forjinn
Forjinn assistants are configurable AI chat interfaces — build one from scratch or connect to an existing OpenAI or Azure OpenAI Assistants service.
Assistants in Forjinn are AI chat interfaces you configure once and share anywhere. You can build a Custom assistant that uses any LLM and tools available in Forjinn, or connect directly to the OpenAI Assistants API or Azure OpenAI Assistants service and manage those assistants from the Forjinn interface.
Assistant types
Forjinn supports three assistant types:
| Type | What it is |
|---|---|
| Custom | An assistant you build entirely inside Forjinn — choose any chat model, attach tools, connect document stores, and write your own instructions. No external assistant service required. |
| OpenAI | A connection to an assistant you've created (or will create) in the OpenAI Assistants API. Forjinn manages the thread and message lifecycle on your behalf. |
| Azure | The same as OpenAI, but pointed at your Azure OpenAI deployment endpoint. Suitable for organizations that require data residency in Azure. |
Create a custom assistant
In the left sidebar, click Assistants. The dashboard lists all assistants in your workspace.
Click Add New. A configuration panel opens.
Select Custom from the assistant type dropdown.
Enter a system prompt in the Instructions field. Describe the assistant's persona, its purpose, and any constraints. You can click Generate to have Forjinn draft instructions based on a description you provide.
Pick a model from the Chat Model dropdown. Any LLM provider configured in your credentials is available — OpenAI, Anthropic, Google, Mistral, and others.
Under Tools, select any tools you want the assistant to use — web search, code interpreter, custom functions, or MCP tools. Tools are listed from the same pool available in Langflow and agent flows.
Under Document Stores, select a document store to give the assistant a knowledge base. The assistant searches the store at runtime to ground its answers.
Click Save. The assistant is ready to test or share.
Connect to the OpenAI Assistants API
Go to Credentials in the sidebar and add a credential of type OpenAI API. Paste your OpenAI API key.
In Assistants, click Add New and select OpenAI as the type.
Choose the OpenAI credential you just created.
Forjinn loads the assistants on your OpenAI account. Select an existing assistant, or click Create New to define one — name, instructions, model, and tools — directly from Forjinn without going to the OpenAI dashboard.
Click Save. Forjinn now manages conversations with this assistant using the OpenAI Threads API.
If you use the OpenAI Assistants API with file uploads (Code Interpreter or file search), you can manage the assistant's vector store and attached files from Settings → OpenAI Files in the Forjinn sidebar.
Connect to Azure OpenAI Assistants
The process mirrors the OpenAI setup, with two differences:
- Add a credential of type Azure OpenAI API instead of OpenAI. You'll need your Azure endpoint URL, API key, and deployment name.
- Select Azure as the assistant type when creating the assistant.
All other configuration options — instructions, tools, document stores — work identically.
Test your assistant
After saving, click the Chat icon on any assistant card to open a test conversation. Messages you send here go directly through the assistant's configured LLM and tools. Sessions are stored in the assistant's chat history.
Share or embed your assistant
Shareable link
Open the assistant's overflow menu and click Share. Toggle Public to generate a URL anyone can open in a browser without logging in to Forjinn.
Embed widget
Click Embed to get a JavaScript snippet. Paste it into your website's HTML and a chat widget appears on the page, powered by your assistant.
To restrict access, assign an API key to the assistant from its Settings panel. Callers must include the key as a bearer token in the Authorization header.
RBAC permissions
Assistant management is controlled by role-based permissions:
| Permission | What it allows |
|---|---|
assistants:view | List and open assistants |
assistants:create | Create new assistants |
assistants:update | Edit existing assistant configuration |
assistants:delete | Remove an assistant from the workspace |
If you can see the Assistants section but can't create or edit, ask your workspace admin to check your role's assistant permissions under Roles in the management section.
Automate workflows with AutoGen in Forjinn
AutoGen lets you build automated multi-agent workflows for task completion, code generation, and research — all visually configured from Forjinn's canvas.
Tools: Extend Your Agents With External Actions
Give your Forjinn agents the ability to search the web, run code, call external APIs, and execute custom JavaScript or Python functions you define.