Platform
Agent runtime
Define an agent once and run it on our side, with tools, memory, step limits and a full log of every step.
https://api.launchad.si/v1/agents/researcher/runs -d '{"message":"Summarise this week in AI pricing"}'No loop to build. Planning, tool calls and retries run on our side.
Every step logged. Follow a run live or read it afterwards.
Report back. A signed run.finished webhook when it is done.
Runs
Start a run,
watch every step
A run returns 202 straight away. Poll it, or stream its steps as server-sent events. Each step records the tool, the input and the result.
Live events GET /v1/runs/{id}/events, resumable.
Child runs Agents can delegate to other agents.
Cancel any time POST /v1/runs/{id}/cancel.
plan find pricing changesweb_search 9 resultsfetch_page acme.dev/pricingnote Pro +$5answer 2 of 3 changed
Safety
Agents that act,
within limits
Agents can work in connected apps. Write actions are capped per run, every action is written to the audit log, and a run can execute as a dry run first.
Dry run See what an agent would do before it does it.
Caps per run Pages, searches, app calls and writes are limited.
Memory sessions Keep context across runs when you want it.
HTTP/2 202{ "id": "run_7f3a…", "status": "queued", "agent": "researcher" }
Also included
The details
that matter
Built-in tools
Web search, page fetch, notes, delegation.
HTTP tools
Your own endpoints as tools, SSRF-safe.
Business apps
Agents act in connected apps such as GitHub or Notion.
Security checks
Agents can check links and IPs mid-run.
Webhooks
Signed run.finished events with retries.
Runs per workspace
Concurrency limits you control.
Related
launchad · getting started
# 1. point your OpenAI client at LaunChad
export OPENAI_BASE_URL=https://api.launchad.si/v1
export OPENAI_API_KEY=$LAUNCHAD_KEY
# 2. call any model, or start an agent run
curl $OPENAI_BASE_URL/agents/researcher/runs \
-H "Authorization: Bearer $OPENAI_API_KEY" \
-d '{"message":"What changed on our competitors' pricing pages?"}'Get started
Two lines of config.
The rest is built.
Keep your OpenAI client, change the base URL and the key. Limits, logs and guardrails are on from the first request.