Phantombuster
Sales IntelligenceLaunch and manage Phantombuster automation agents (phantoms) and retrieve their scraped results. Agents start a phantom, poll its run to completion, pull back the collected data, and monitor organization quotas.
What This Integration Enables
Agents use Phantombuster to fold a scraping or automation run into an orchestrated workflow instead of running it as a manual, out-of-band task. Launch Agent starts a phantom with an optional argument object that overrides the phantom's saved input for that run, and returns a containerId. Get Agent Output and Get Container poll the run until it reports finished, and Get Container Result Object returns the structured data the phantom collected, ready to hand to the next step. Around the core run, List Agents and List Agent Containers let an agent discover phantoms and find past runs, Abort Agent stops a run that is stuck, and Get Organization Resources reports execution-time and storage quotas so an agent can check headroom before launching. The result is that scraped data becomes an input to a flow, not a file someone has to move by hand.
Without FlowRunner
With FlowRunner
Use Case Scenarios
Scheduled Lead Collection
On a schedule, an agent checks Get Organization Resources for available execution time, then calls Launch Agent to run a lead-scraping phantom with the week's search input. It polls Get Agent Output until the run finishes, reads the results with Get Container Result Object, and appends each collected lead to a sheet and the CRM. The team starts the week with a fresh, structured list instead of a scraping job someone had to remember to run.
Poll and Route Results
An agent launches a profile-scraping phantom and needs to react the moment it completes. It polls Get Container on the returned containerId until the status is finished, pulls the result object, and routes each record to the right downstream system based on its content. The scraping run and the routing logic live in one workflow rather than two tools stitched together by a person.
Quota Guard and Alert
An operations agent runs Get Organization Resources on a schedule and watches execution-time and storage quotas. When a quota is close to its limit, it alerts the owner so runs do not fail silently mid-campaign. If a run gets stuck, the agent calls Abort Agent to free the container rather than letting it consume quota. The team keeps its automation capacity under control without watching a dashboard.
Human-in-Loop Highlight
A phantom acts against real accounts, often with a session cookie, and that carries a responsibility a data lookup does not. Before an agent launches a phantom that will operate against live accounts, it pauses and asks a human to approve the run and its input. This is FlowRunner's human-in-the-loop applied to the moment that matters: an agent that knows when to stop and ask for help rather than acting on a target list unattended. The agent messages the owner through Slack: "I am about to launch [Phantom] against [target] with this input. Approve the run, or adjust the input before it touches live accounts." The owner reviews the phantom and its scope and approves, and the automation runs against targets a person signed off on. The agent handles the launch, polling, and result collection; the human owns the decision to act on real accounts.
Agent Capabilities
9 actionsAgents
5- List Agents List all agents (phantoms) in your organization, each with its id, name, script id, launch settings, and last-run metadata. Use it to discover which phantom to run and to feed an agent id into Launch Agent.
- Get Agent Fetch a single agent by its id, including its configuration, launch arguments, script id, and last-run status. Use it to inspect a phantom's expected argument shape before launching it.
- Launch Agent Launch an agent to run in a new container. Optionally pass an argument object to override the saved input for this run, a bonus argument merged on top, and saveArgument to persist it as the default. Returns the containerId. Use it to start a phantom, then poll for completion.
- Get Agent Output Return the current or latest run output for an agent: console and progress output, run status, progress percentage, and the resultObject when available. Use it to poll an in-progress launch until its status is finished.
- Abort Agent Abort the currently running container(s) of an agent, stopping the automation immediately. Returns the number of containers stopped. Use it to cancel a run that is stuck or no longer needed.
Containers
3- Get Container Fetch a single container by its id. A container is one run of an agent; the response returns its status, launch metadata, and timing. Use it to poll a specific run started by Launch Agent until it finishes.
- Get Container Result Object Return the structured JSON result object produced by a container run: the scraped or automated data the phantom collected. Call it with the containerId once the run reports finished. This is the key operation for retrieving results.
- List Agent Containers List all containers (runs) for a given agent, most recent first, with each run's id, status, and timing. Use it to find past runs and retrieve their results.
Organization
1- Get Organization Resources Return your organization's resource usage and limits, including execution time, storage, and email quotas. Use it as a connection check and to confirm headroom before launching agents.
Start building with Phantombuster
$100 in credits. No card required. Connect in minutes.