z3nDash

Stores schedules and runs local tasks. Each task carries an executor, a path, arguments, a schedule and an overlap policy; output streams live while it runs.

List

Tasks are read from the schedules table. The list shows the name, the state and the main run parameters.

Editing

A task carries a name, an executor, a path to a script, file or task, arguments, an enabled flag, a schedule, an overlap policy and a payload. Executors: python, node, ts-node and npm run scripts of their runtimes; exe, cmd, bat, bash and ps1 run executables and shell scripts; csx runs a C# script; csx-internal runs one with access to z3nDash internals; xml plays a ZennoPoster template — the built-in runtime executes the XML graph directly, with the browser chosen on the task card; internal runs a built-in z3nDash task.

Schedule

The UI builds on-demand runs, daily runs, runs on days of the week, runs on a day of the month, and minute intervals. The result is stored as cron, interval_minutes and fixed_time.

Overlap

skip — do not start a new instance while one is active; parallel — allow several instances up to max_threads; kill_restart — stop the active ones and start fresh.

Control

Run, Stop, Restart, enable or disable the schedule, delete and duplicate, open the file or its folder, open a terminal, check and install Python and Node dependencies.

Output

Output of the active run streams over SSE. Parallel instances each get their own runId.

Shows task state pulled directly from the registered ZP node services. The shared log history and the HTTP traffic window both open from here.

Data source

z3nDash stores only the node addresses. Task state is requested from each node directly, so nothing is cached behind your back.

Task list

Filters by name, state, run mode and tags. The chosen filters are kept in the UI state.

Control

Commands go straight to the selected node: start, stop, interrupt, apply or update settings, add or set tries, set threads, clear done, clear fails, execution settings, scheduler settings.

Logs window

allLogs opens the shared history across every registered node or one chosen machine, with filters by project, level, thread and module, and full-text search.

Traffic window

traffic opens the HTTP traffic window. Nodes and the local runtime write traffic to a file and the window reads its tail; the former separate HTTP page and its store are gone.

Surfaces data from the external ZennoBoxer API, alongside local process operations such as reading uptime by PID and terminating a process tree.

Connection

The address comes from ApiConfig.ZbHost, with a fallback when it is empty. The token is passed upstream in the Api-Token header.

Data

The page proxies ZennoBoxer data — profiles, proxies, instances and threads. Which fields appear depends on what the external API returns.

Local operations

Read process uptime by PID, and terminate a whole process tree by PID.

Opens an HTTP archive locally and lets you filter, replay and export the requests inside it. The file is parsed in the browser and kept in IndexedDB, so it survives a reload.

Data source

A HAR file is opened locally: the page reads log.entries and parses it in the browser. The page itself has no server routes. Entries can also be added by importing cURL, with no file at all.

Storage

The loaded archive is kept in IndexedDB and survives a page reload. Clear HAR wipes both the list and the stored copy. Nothing reaches the z3nDash database.

List

Filters by method, status, URL and project, plus a limit on how many records are shown. Refresh re-applies the filters to data already loaded and never touches the network; Reset clears the filters.

Replay

A selected request is sent through the same replay listener the HTTP page uses. Method and URL can be edited before sending, and the response can be copied.

Code

For a selected record the page generates stubs: API Skeleton, API Example, HttpClient, ZP7, Hybrid, Python, TypeScript and cURL.

A local JSON viewer and editor. Works entirely on the machine, with security badges over the document and optional AI-assisted analysis.

Document

Paste JSON, format it, walk the value tree, search, collapse branches, copy a value or its path, filter fields.

Security badges

Values that look like secrets or sensitive data get marked. This is a heuristic highlight, not a security check — treat it as a hint.

Replay

The document can be used as the body when preparing an HTTP replay.

AI analysis

Optional analysis goes to OmniRoute; results are cached in the ai_json_cache table.

A set of local text transforms. Nothing is sent anywhere — the conversions run in the dashboard itself.

Tools

URL encode and decode, Base64 encode and decode, C# string escaping, JSON string escaping. Everything runs in the browser and needs no database.

Keeps named text templates for quick copying, arranged as a tree. Useful for anything typed often enough to be worth storing once.

Capabilities

Browse the template tree, select and copy content, create and edit entries, store a path, a title and the text itself.

Storage

Everything lives in the clips table, so templates survive restarts and are shared by every page that reads them.

A standalone viewer for a chosen SQLite file. It is not the z3nDash datastore — application data lives in PostgreSQL.

Capabilities

Open a database file, list its tables, run a query, change a value, delete a row.

Scope

This screen never switches the main z3nDash database. It works only on the file passed to the viewer API — application data stays in PostgreSQL.

Collects and displays Web3 balances for the addresses stored in the database. Refreshes run in the background and can be triggered over the API.

Refresh

POST /treasury/update starts a background refresh with MaxId, minValue and a concurrency between 1 and 20. Progress is available from GET /treasury/status.

Presentation

Total value, balances by account and chain, top tokens, distribution across chains and a portfolio summary. Which chains appear is decided by the columns of the actual _treasury table.

AI

Portfolio analysis goes through OmniRoute; the result is kept in the Treasury AI cache.

Captures a diagnostic snapshot of the current Windows system and writes it to a file, so two moments in time can be compared later.

Capture

Capture Now collects a snapshot on the machine z3nDash runs on. It includes the sections the backend actually managed to gather: processes, network and system data.

Files

Load a snapshot from a file, save one to the database, open a stored snapshot or delete it. Snapshots live in the system_snapshots table.

AI audit

An optional audit sends the snapshot to OmniRoute; the result is cached in system_snapshot_ai_cache.

Manages the local z3nDash configuration: database connection, server and log paths, integrations, and the process-memory watchdog.

Server status

Configuration state, database connection, dashboard port, log and report folders, and the selected database mode. Loaded from GET /config/status.

Database

Two modes: PostgreSQL with a single connection string, or SQLite with a path to a file. For PostgreSQL the schema comes from Search Path. The connection string can be copied with the button beside the field.

Logs & server

Addresses for logs and HTTP traffic, the dashboard port, and the working folders.

OmniRoute

OmniRoute is the only AI provider. The field holds its URL, http://localhost:20128 by default. Validate & Save AI stores the address and checks that /v1/models answers.

Security

The jVars section stores encrypted local variables and the path to the jVars file.

Memory watchdog

Root process name, an on/off switch, a memory limit in MB and a check interval. The watchdog counts the root process together with its child tree and, once a non-zero limit is exceeded, terminates the tree.

Renders the Markdown files of a docs vault as a linked graph. Choose folder picks the vault; by default it is the one bundled with the application.

Choosing a vault

Choose folder opens a system dialog for picking the vault. By default it is the docs-vault bundled with the application.

Generation

POST /docsVault/generate reads the Markdown files and builds links from wiki references. A relative path is resolved against the application folder.

Viewing

The last graph built is returned on request; before the first generation an empty template opens.

Export

The graph can be downloaded as a standalone HTML file.

Builds a graph of C# code. Choose DLL picks a compiled assembly, or point it at a folder of sources; relationships between types and members are then explored visually.

Input

Choose DLL opens a system dialog for picking a compiled assembly; Generate then builds the graph. A folder of DLLs or of C# sources works too — bin and obj are skipped when reading sources.

Result

The generated HTML is held in process memory and the last graph built is returned on request. After a restart the graph has to be built again.