Research Agent
Custom Agent Runtime
The presentation and operating layer for AI agentsGive your agents
Operator preview · Client view and customer branding planned.
Give your agents
a stage.
Turn the agent system you already built into a visible experience for activity, runs, approvals, artifacts, health and client interaction.
Keep the runtime. Upgrade the experience.Operator preview · Client view and customer branding planned.
Northstar AIDemo dataWorkspace / active runSample scenario
Competitive Landscape Research
RResearch AgentCustom Agent RuntimeRunning
run.startedGathering source material
Health: unknown · No signal in this sampleNo live connection
Agents. In their element.Real signals. Explicit boundaries.
What is Stageframe?
Stageframe is the presentation and operating layer for AI agents. It presents reported work from the agent systems you already use and turns runtime activity into understandable workspaces for operators, clients and stakeholders. Stageframe does not replace your agent runtime.
An AI agent dashboard for reported work brings run monitoring, human-in-the-loop attention and output provenance into one understandable view. The operator experience is implemented in a local connected preview. The public site shows deterministic demo data. Selected client access, delivery and customer branding remain planned.
Stop handing clients the backend.
A workflow editor explains configuration. A log explains an event. A client needs to understand the work. Keep the runtime. Upgrade the experience.
The same sample work / backend
Competitive Landscape Research
Three signals, recorded by the runtime. Watch them become a workspace.
- Waiting in demo
run.started14:00 UTC · demo_started - Waiting in demo
artifact.created14:05 UTC · demo_output - Waiting in demo
approval.requested14:06 UTC · demo_review
Source timestamps stay fixed. Playback compresses the sample; it makes no network request.
The same sample work / StageframeNorthstar AIDemonstration data
Northstar AIDemonstration dataRun
No run reported
The first run.started signal identifies the agent and its work. Nothing has been reported yet.
Activity history / UTC
No events reported yet. The next demo signal creates a visible run.
Artifact
No artifact reported. A started run is not an output.
Attention
No human decision reported in this sample state.
Explore the sample
Inspect the run, artifact metadata or review request. These controls reveal demo context and send no runtime action.
See the work. Step in when it matters.
Northstar AIDemonstration dataRun
Competitive Landscape Research
Running
Running
Activity history / UTC
- Research started
run.startedGathering source material
Artifact
No artifact reported. A started run is not an output.
Attention
No human decision reported in this sample state.
Explore the sample
Inspect the run, artifact metadata or review request. These controls reveal demo context and send no runtime action.
An active agent. A named piece of work.
run.started reports work underway. No percentage or estimated finish is invented.
Each question selects a deterministic sample. Completed and failed outcomes use separate run identities; they are not successive states of one run. Manual selection stays paused until you choose Play.
Your runtime. A clearer experience.
Your runtime
Reasoning · Tools · Execution→Stageframe
Reported state · Output · Attention→Operator / client
Operator preview · Client view proposed
Reasoning · Tools · Execution→Stageframe
Reported state · Output · Attention→Operator / client
Operator preview · Client view proposed
Your runtime
Builds and runs the agents.
Reasoning, model selection, tools, orchestration and action enforcement remain in the system you connect.
Stageframe
Makes the work understandable.
Reported activity, attention and output become an operator experience. Selected client delivery remains proposed.
Make the handoff part of the product.
Give the operator depth to supervise the work. Plan a selected client experience around status, output and the next decision. Keep credentials and operational permissions with the builder.
Actual product shellDemo data · Local development
Open full-size workspace image ↗ (new tab)Start with the signals you already have.
- 01 / Report activity
Your system sends the signals.
Map agent identities, run events, output metadata, review requests and health into the V1 event contract.
- 02 / Present work
See what needs you.
Bring activity, attention and artifacts into a consistent workspace. Execution stays in your runtime.
- 03 / Select visibility
Plan the client experience.
Agree what a client should see. Client access and delivery need separate implementation.
Proposed
Designed for the stack you already use.
A generic event architecture, with availability stated before the provider name implies support. No generally available native connection is verified.
Webhooks & APIPreviewn8nPlanned · Target integrationLangGraphPlanned · Target integrationEventus EngineDesign Partner #0 · Native integration planned
Review every connection path →One project. Ongoing building. Client delivery.
Choose the engagement you want to discuss. These prices test interest; scope and access are reviewed individually before any charge.
Project Pass
Early access$199one-time
One client. One project. 30 days.
For a client delivery, proof of concept, pilot or short engagement.
Pricing hypothesis · scope agreed individuallyBuilder
Early access$149/month
For builders using Stageframe continuously.
For independent builders, consultants and small teams with ongoing implementations.
Pricing hypothesis · scope agreed individuallyAgency
Agency launch focus$299/month
For teams delivering agent systems to clients.
For AI agencies, automation agencies and consultancies making delivery a repeatable practice.
Pricing hypothesis · scope agreed individuallyExplicit trust. Constrained authority.
Workspace scope, least privilege, credential separation and authenticated events are the baseline. Local tests verify constrained paths; production customer authentication and access remain release gates.
Understand the boundary before connecting.
Who is Stageframe for?
The primary launch focus is AI agencies, automation agencies, agent builders and consultants already delivering systems. SaaS teams and internal AI teams can also review fit.
What does Stageframe show?
Reported activity and run state, output metadata, attention and explicit health signals. The operator preview is implemented locally; selected client access and delivery remain planned.
Does Stageframe build or run my agents?
No. Your connected runtime owns reasoning, model selection, tool use and execution. Stageframe presents reported work and bounded interaction surfaces.
Which systems can I connect?
The signed V1 event path and Eventus reference contract are tested locally. A connection needs mapped run, artifact, attention and health telemetry. An OpenAI-compatible endpoint alone does not provide those events.
How does Stageframe connect?
A runtime-side sender maps stable identities and work events to the signed V1 contract. The receiver validates schema, signatures and scope. The local Webhooks & API path is a preview, not a generally available hosted integration.
Can I use Stageframe with my existing agent stack?
Stageframe is designed to preserve your runtime. A connection needs verified telemetry mapping, scoped credentials and an adapter; model API compatibility alone is not enough. Named provider targets do not imply native support.
What can clients see?
The current client view is a proposed preview of selected activity and output. Authenticated client access, public share links and file delivery are not connected.
Can clients approve actions?
The public demo only illustrates review. A bounded operator approval return is tested in the local slice. The runtime enforces the action; general client approvals and permission expansion are not available.
How does Founder Access work?
Send a request describing your existing system. Requests are stored for individual review; connection fit and any onboarding plan still need agreement. Submission does not grant access or guarantee a response time.
What does it cost?
Project Pass, Builder and Agency are early-access pricing hypotheses. There is no checkout or automatic access grant; scope and commercial terms must be agreed before paid onboarding. See the pricing page for current amounts.
For builders already delivering agent systemsThe work is already happening.
Make it visible.
Describe your runtime, the output it creates and who needs to understand it. Your request is saved for individual review.
See Stageframe