Using the Web App
This walkthrough covers the complete apps/web journey: sign up, create a Lens, build a Workflow, connect an AI lenser, and execute your first run — all from the browser.
Prerequisites: Complete Installation and have the app running at http://localhost:3000.
Step 1 — Sign up or sign in
Open http://localhost:3000 and create a local account. After registering you will land on the onboarding flow.
Complete the basic profile setup:
- Choose a handle (your
@usernameon the platform) - Set a display name
- Optionally add a bio and avatar
Your profile is now accessible at /lenser/yourhandle.
Step 2 — Create your first Lens
Navigate to Lens Library → New Lens (or /lenses/new).
A Lens is a task specification. Fill in:
- Title — a short human-readable name (e.g.,
Explain a Concept) - Template body — the task text, with optional
[[parameter]]placeholders:
Explain [[concept]] to a complete beginner in under 100 words.
Use simple analogies and avoid jargon.- Parameters — add the
conceptparameter as typetext - Tags — optional topic labels for discovery
- Visibility — start with
privatewhile drafting
Click Save Draft, review it, then click Publish.
Published Lenses are frozen. To revise, create a new version from the lens detail page.
Step 3 — Build a Workflow
Navigate to Workflows → New Workflow (or /workflows/new).
A Workflow is a DAG (directed acyclic graph) of Lens nodes. For your first Workflow, build a two-step pipeline:
Add nodes
- Click Add Node and select the Lens you just created (
Explain a Concept) - Add a second Lens — use the built-in Summarize Lens from the public library
Connect the nodes
Drag from the output handle of node 1 to the input handle of node 2. A panel opens asking which output field maps to which parameter of the next node.
Map: node1.output → [[text_to_summarize]]
Set context inputs
At the top of the Workflow, set the root node's input:
concept = "recursion"Click Save Workflow.
Step 4 — Connect a lenser
Before executing, you need a lenser (AI Lenser) connected to your account.
Go to AI Workspace → Runners → Connect Lenser.
Choose a type:
| Option | When to use |
|---|---|
| Ollama (local) | No API key needed; runs on your machine |
| OpenAI | Requires OPENAI_API_KEY in your environment |
| Custom HTTP | Any endpoint that accepts a Lens and returns output |
For a quick local run, choose Ollama and enter llama3.2 as the model.
Click Connect. The lenser appears in your runners list with status active.
Step 5 — Execute the Workflow
Open your Workflow and click Run.
A run panel opens showing:
- Run status — live updates as the run progresses
- Node status — each node shows
pending → running → completed - Streaming output — the AI's response appears in real-time as it is generated
For a typical two-node Workflow on a local model, expect the run to complete in 5–30 seconds depending on your hardware.
Step 6 — Inspect the results
After the run completes:
- Click any node to see its output, input parameters, duration, and cost
- Click Run Log to see all events in order
- Use the Provenance tab to trace how data flowed between nodes
If a node failed, you will see the error message and can retry directly from the UI.
Step 7 — Publish and share
When your Workflow is ready to share:
- Open the Workflow and click Publish
- Set visibility to public or unlisted
- Copy the shareable URL:
/workflows/yourhandle/your-workflow-slug
Other Lensers can fork your Workflow to build their own version.
Web app sections at a glance
| Section | URL | What you do there |
|---|---|---|
| Home feed | / | Discover trending Lenses and Workflows |
| Lens Library | /lenses | Browse, create, and manage Lenses |
| Workflows | /workflows | Create and run multi-step pipelines |
| AI Workspace | /lenser/yourhandle/ag/overview | Connect runners, manage teams, view run history |
| Profile | /lenser/yourhandle | Your public profile and lens collection |
| Explore | /explore | Search Lensers, Lenses, and Workflows |
What to do next
- Create a Lens (detailed) — Full Lens editor walkthrough
- Create a Workflow (detailed) — Multi-node Workflow guide
- CLI Getting Started — The same journey from the terminal
- Agent Teams — Group runners for complex multi-agent pipelines
- Executions — Deep dive into how runs work