Surfaces
Agent Studio
Prototype, configure, and test Allternit agents in the platform shell
Documentation Index
Fetch the complete documentation index at: /llms.txt
Use this file to discover all available pages before exploring further.
Prototype, configure, and test Allternit agents in the platform shell
https://ai.allternit.com/shell?view=agent-studio
open('agent-studio');
| Tab | Purpose |
|---|---|
| Agent | Set name, description, system prompt, model, temperature, and max tokens |
| Tools | Toggle tools the prototype may call |
| Run | Send a test message and inspect the raw response |
| Field | Type | Default | Description |
|---|---|---|---|
name | string | My Agent | Agent display name |
description | string | "" | What the agent does |
system_prompt | string | You are a helpful Allternit agent. | System instructions |
model | string | allternit-large | Brain model identifier |
temperature | number | 0.7 | Sampling temperature |
max_tokens | integer | 4096 | Token limit |
tools | string[] | [] | Enabled tool IDs |
| Tool ID | Name | Description |
|---|---|---|
web_search | Web Search | Search the web for current information |
bash | Bash | Run shell commands in a sandbox |
text_editor | Text Editor | Read and edit files |
computer_use | Computer Use | Control a desktop environment |
mcp_registry | MCP Registry | Use registered MCP tools |
| Method | Path | Description |
|---|---|---|
POST | /api/v1/agents/prototype | Run a prototype agent with the configured settings |
POST | /api/v1/agents | Save the configured agent (Phase 2) |
curl -X POST https://api.allternit.com/api/v1/agents/prototype \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
-d '{
"name": "Research Agent",
"description": "Performs deep technical research.",
"system_prompt": "You are a research assistant. Cite sources.",
"model": "allternit-reasoning",
"temperature": 0.5,
"max_tokens": 4096,
"tools": ["web_search", "mcp_registry"],
"messages": [{"role":"user","content":"Summarize local model quantization."}]
}'
{
"role": "assistant",
"content": "Local model quantization reduces...",
"tool_calls": [
{
"id": "call_1a2b...",
"type": "web_search",
"arguments": { "query": "local model quantization techniques" }
}
]
}
{
"name": "Research Agent",
"description": "Performs deep technical research.",
"system_prompt": "You are a research assistant. Cite sources.",
"model": "allternit-reasoning",
"temperature": 0.5,
"max_tokens": 4096,
"tools": ["web_search", "mcp_registry"],
"messages": [{"role":"user","content":"Summarize local model quantization."}]
}
| HTTP | Code | Meaning |
|---|---|---|
| 200 | — | Prototype completed |
| 400 | invalid_request | Missing required fields or invalid tool IDs |
| 401 | unauthorized | Missing or invalid bearer token |
| 500 | prototype_failed | Brain or runtime error during the prototype run |