Recruiting design partners for new verticals — open a new industry, get the platform at half price
Nexevo.AI
Log in

Quickstart

Three steps to your first agent run. You'll send a task and read the streamed result — the platform handles everything else.

1. Get an API key

In the app, open Settings → API keys and create a key. Keys are workspace-bound (sk-ws-…). Or create one over the API:

POST/v1/workspaces/{workspace_id}/api-keys —

create a new API key bound to this workspace.

bash
export NEXEVO_API_KEY="sk-ws-..."
export NEXEVO_WORKSPACE="00000000-0000-0000-0000-000000000000"

2. Send a task

That's the whole integration. One field — task. The platform picks the model, calls whatever tools the task needs, and streams the result.

bash
curl https://nexevo.ai/v1/workspaces/$NEXEVO_WORKSPACE/agent/runs \
  -H "Authorization: Bearer $NEXEVO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"task": "Find the 3 largest cross-border payment providers in SE Asia and summarize each in one line."}'
# -> {"run_id": "…", "trace_id": "…"}

Returns {run_id, trace_id}. The run is now executing — you don't need to stay connected.

3. Read the result

bash
curl -N https://nexevo.ai/v1/workspaces/$NEXEVO_WORKSPACE/agent/runs/$RUN_ID/attach \
  -H "Authorization: Bearer $NEXEVO_API_KEY"

Each data: line is an event — step_start, tool_call, tool_result, llm_delta (streamed text), and finally run_end (with final_text). The stream ends after the terminal frame.

That's it — you've completed a full integration. The platform auto-handled model selection, web search, and result synthesis. You configured nothing beyond the task itself. For more scenarios (OpenAI SDK, knowledge-base Q&A, scheduled reports), see the Cookbook.

Next steps