Skip to main content

FlowBase CLI

flowbase is the official command line client for the FlowBase OpenAPI. It manages apps, objects, fields, views, custom pages, workflows, approval flows, dashboards, business records, roles, and configuration versions across the three FlowBase environments.

FlowBase is a low-code platform. Start with flowbase guide (offline): discover apps and objects, read fields, then configure and verify. Use flowbase guide --format json for structured guidance.

1. Install and connect

python -m pip install ./packages/flowbase-cli

Connect to an environment by its site Base URL and a FlowBase API key:

export FLOWBASE_HOST=https://flowbase-test-a.deepflowagent.com
export FLOWBASE_API_KEY=fbk_xxx
Setting Source
Host --host > FLOWBASE_HOST > http://flowbase.localhost:5001
API key --api-key > FLOWBASE_API_KEY

The CLI appends /api/openapi/v1 automatically and sends Authorization: Bearer <key> plus an X-FlowBase-Channel: cli header. It never stores keys on disk and never prints the full key in errors.

2. Environments

FlowBase has three environments, selected only by the Host — never by a query, body, or extra header:

Environment Host
Production https://flowbase.deepflowagent.com
Test A https://flowbase-test-a.deepflowagent.com
Test B https://flowbase-test-b.deepflowagent.com

Most commands run on every Host; roles is production-only and changes commit/release is test-only.

3. Mental model

3.1 Metadata vs records

  • Metadata (metadata scope): apps, objects, fields, views, pages, workflows, approvals, dashboards, roles, versions.
  • Business records (records scope): rows inside an object.

Metadata changes are drafts against the current environment's configuration repository. They do not affect the running front-end until they are committed and released.

3.2 Draft → commit → release

metadata write commands  →  draft in the current environment
changes commit           →  snapshot draft into a new version (test env)
changes release          →  publish the current test version to production
changes direct-release   →  production hotfix: preflight + confirmation, then publish drafts

Most update commands accept an optional baseRevision to guard against concurrent edits; the server returns 409 if the draft moved since your last read.

3.3 Scope and confirmation

Each operation is in metadata:read, metadata:write, records:read, or records:write. Destructive commands (delete, restore, set-enabled, version operations) ask for confirmation in a terminal; in automation you must pass --yes.

4. Command groups

flowbase
├── docs            list | show | example | operations   Offline endpoint catalog
├── guide           Offline product and task guide (--format markdown|json)
├── organization    employees | departments
├── apps            list | get | create | update | set-enabled | delete | navigation
│                   add-objects | remove-object | set-group | reorder-objects
│                   add-dashboards | remove-dashboard | set-dashboard-group | reorder-members
├── groups          create | reorder | update | delete    App groups (under an app)
├── objects         list | get | create | update | delete | restore
│                   cli-edit-policy get | update
├── fields          create | update | delete | restore    Fields under an object
├── views           list | create | update | reorder | set-enabled | delete
├── workflows       list | get | create | update | reorder | set-enabled | delete | executions
├── approvals       list | get | create | update | reorder | set-enabled | delete
├── pages           list | options | get | create | update | set-default | clear-default | delete | restore | evaluate
├── dashboards      list | get | create | update | copy | delete | run
│                   record-view-options | record-view-preview | records
├── changes         commit | release | direct-release
├── versions        list | show | releases                Read-only history
├── environments    restart                              Destructive test-env reset
├── records         list | get | create | update | delete | upload-attachment | download-attachment
└── roles           list | options | get | create | update | delete   (production only)

Global flags: --host, --api-key, --pretty/--no-pretty, --raw. They may appear before or after the subcommand (flowbase apps list --no-pretty is valid).

5. Discovering request bodies

JSON writes accept --json or --json-file. The CLI is self-describing:

flowbase docs list                       # endpoints + hasRequestBody/requiresConfirmation flags
flowbase docs show dashboards.update     # metadata + representative request body
flowbase docs example dashboards.update --output dashboard.json
flowbase dashboards update dsh_xxx --json-file dashboard.json

flowbase docs example <key> prints a representative body template for that endpoint so you can edit it instead of writing from scratch.

Closed-domain catalogs

Complex bodies reference closed sets that are not guessable from examples. Query them offline:

flowbase docs catalogs                    # list available catalogs
flowbase docs catalogs authoring          # required fields, draftRevision sources, common mistakes
flowbase docs catalogs field-types        # types, constraints, filter operators, record values, create examples
flowbase docs catalogs view-filters       # view filter operators and value shapes (not page-rules)
flowbase docs catalogs record-values      # how to write records.values by field type
flowbase docs catalogs workflow-nodes     # node types with in/out edge rules
flowbase docs catalogs permissions        # function permission points + data scopes
flowbase docs catalogs dashboard-filters  # filter types + date-range presets
flowbase docs catalogs page-components    # Page component shapes, placement, fields, and enums
flowbase docs catalogs page-rules         # Page condition operators and action shapes
flowbase docs catalogs forms              # custom form fields, rules, appearance, and settings
flowbase docs catalogs id-prefixes        # resource ID prefix meanings
flowbase docs catalogs icons              # icon guidance + examples

Read authoring before composing any write. It lists the mistakes that most often make generated commands fail: missing description/type, select options without id+lifecycle, defaultValue: null (omit the key instead), boolean view filters using eq, records.list being a column projection, pages.list.pages being a page count, and objects.get not returning draftRevision.

The read → modify → write loop

Most update commands replace a full snapshot and accept an optional baseRevision for optimistic locking. Check the relevant offline catalog for its exact update envelope and read-only response fields. The safe pattern is:

flowbase dashboards get dsh_xxx > current.json      # authoritative snapshot
# Keep only label/description/icon/filters/components; use draftRevision as baseRevision.
# Edit the desired fields while preserving all other components and filters.
flowbase dashboards update dsh_xxx --json-file current.json

Do not compose update bodies from scratch when an existing resource exists. Start from get, then convert its read projection to the update envelope documented by the relevant catalog.

6. Recipes

6.1 Create an object with a field and a record

flowbase docs catalogs authoring
flowbase docs catalogs field-types
flowbase objects create --json '{"label":"Customer","key":"customer","description":"Customer records","icon":"IconUsers"}'
flowbase fields create obj_xxx --json '{"label":"Name","key":"name","description":"Customer name","type":"text.singleLine","typeVersion":"v1","permission":"required","constraints":{"maxLength":120}}'
flowbase records create obj_xxx --json '{"values":{"name":"Acme Corp"}}'
flowbase objects cli-edit-policy get obj_xxx
flowbase objects cli-edit-policy update obj_xxx --json '{"baseRevision":4,"policies":{"fld_xxx":"user_approval"}}' --yes

records create returns {recordId}; use it with records get.

To add existing objects to an application, pass their IDs with --object-ids:

flowbase apps add-objects app_xxx --object-ids obj_a,obj_b

Use a comma-separated list for multiple objects. Object IDs are not positional arguments. Run flowbase docs show apps.objects.add for the endpoint details.

Record values keys are field keys (name), not internal IDs (fld_...). IDs are obj_/fld_/dsh_/wfl_/pag_/reg_/cmp_-prefixed; see docs catalogs id-prefixes.

6.2 Edit and publish configuration on a test environment

export FLOWBASE_HOST=https://flowbase-test-a.deepflowagent.com

flowbase dashboards create --label "Sales board" --key sales_board
flowbase docs example dashboards.update --output dashboard.json   # fill filters/components
flowbase dashboards update dsh_xxx --json-file dashboard.json
flowbase apps add-dashboards app_xxx dsh_xxx
flowbase apps reorder-members app_xxx --json '{"items":[{"type":"object","id":"obj_xxx"},{"type":"dashboard","id":"dsh_xxx"}]}'
flowbase changes commit --message "Add sales dashboard" --yes
flowbase changes release --message "Ship sales dashboard" --yes

Dashboard component code runs server-side in a sandbox and returns ECharts config / metric / table / markdown. Components and filters are validated on save; use flowbase dashboards get <id> to discover component IDs, then flowbase dashboards run <dashboard_id> <component_id> to inspect ok/output/error/logs in a read-only draft preview (requires metadata:read and records:read). Visual chart rendering still requires the web UI.

App sections and record-view components:

flowbase apps get app_xxx
flowbase docs catalogs app-sections
flowbase apps navigation app_xxx --json '{"type":"appSection.create","name":"Sales"}' --yes
flowbase groups create app_xxx --name Orders --section-id sec_xxx
flowbase apps add-objects app_xxx --object-ids obj_xxx --section-id sec_xxx
# add-dashboards and add-iframes also accept --section-id.
flowbase apps navigation app_xxx --json '{"type":"appMember.move","id":"obj_xxx","sectionId":"sec_xxx","groupId":"grp_xxx","expectedSectionId":"sec_xxx","expectedGroupId":null}' --yes

flowbase docs catalogs dashboard-record-views
flowbase dashboards record-view-options
flowbase docs example dashboards.record-view-preview --output preview.json
# Replace placeholder IDs and configure component/filters before previewing.
flowbase dashboards record-view-preview dsh_xxx --json-file preview.json
flowbase dashboards records dsh_xxx dcm_xxx --page 1 --query Orders --json '{"filterValues":{}}'

apps get preserves section definitions and group/member sectionId. Navigation supports section creation, rename, empty-section deletion, full section ordering, group moves and member moves. Moves require explicit source and target positions; use JSON null for unassigned/ungrouped. See app-sections for all six request bodies. Navigation requires confirmation (--yes for automation), and writes only a draft.

Record-view previews and queries require both metadata:read and records:read, use the current Host's records and draft configuration, and return mode: draft-preview plus draftRevision. They use API-key visibility rather than an employee's workspace identity; hidden fields stay excluded and hidden filter fields are rejected. Page size is configured on the component (10/20/50). Search is ignored when showSearch is false. dashboards run is for code components.

For dashboards update, construct the body from the latest get: preserve label, description, icon, filters, and all components; set baseRevision to the returned draftRevision. Omit response-only fields. Keep existing component IDs and recordView unchanged when editing other properties; omit IDs on new components. Omitting components or filters clears them.

6.3 Production hotfix

export FLOWBASE_HOST=https://flowbase.deepflowagent.com
flowbase changes direct-release --message "Fix customer field" --yes

direct-release always runs a server-side preflight, prints the current version, diff summary, checks and risk, and binds a one-time confirmation token. --yes accepts that reviewed result; it never skips validation.

On production, records.* sees released metadata only. Creating records on a brand-new object before direct-release returns 404 内容不存在或已被删除. Finish metadata, release, then seed.

6.4 Workflows, approvals, and pages

A workflow graph must satisfy structural rules: exactly one trigger node, at least one end node, a DAG, and valid per-type edges. See docs catalogs workflow-nodes before composing nodes/edges.

Workflow conditions use fieldId for current-object fields and add relationFieldId for a one-hop related field. Field-value references use valueRef plus optional valueRefRelationFieldId. An updateRelated node selects the target record with its required relationFieldId; its assignment fieldId values belong to that related object. Use workflows list to retrieve the current object fields and the available relations with target fields.

Approval flow graphs are separate metadata: exactly one trigger, at least one approval node, both approvedEnd and rejectedEnd, and dual outcome edges from each approval node. Employee task decisions are not available via API key / CLI. See docs catalogs approval-nodes.

flowbase docs catalogs workflow-nodes
flowbase workflows create obj_xxx --json '{"label":"Set owner","enabled":true,"trigger":{"event":"create","match":"all","conditions":[]},"nodes":[],"edges":[]}'
flowbase workflows update obj_xxx wfl_xxx --json-file workflow.json
flowbase workflows executions obj_xxx wfl_xxx --result failure

flowbase docs catalogs approval-nodes
flowbase docs example approvals.create --output approval.json
flowbase approvals create obj_xxx --json-file approval.json
flowbase approvals set-enabled obj_xxx apr_xxx --base-revision 4 --enabled --yes

flowbase docs catalogs page-components                  # all component shapes and placement rules
flowbase docs catalogs page-rules                       # conditions, operators, and actions
flowbase pages options                                  # objects, fields, and lookup constraints
flowbase docs example pages.create --output page.json   # create-page layout template
flowbase pages create --json-file page.json
flowbase pages set-default pag_xxx --base-revision 4 --yes

# Detail pages support detailSummary, stageBar, tabs, relatedSummary,
# attachmentCard, noticeBlock, relatedList, and approval components.
# A placed field uses sourcePath=[] for the Page object, or 1-2
# {relationFieldId,targetObjectId} hops for a related object field.
# Component visibility uses a page-rules setComponentVisibility action
# targeting the stable cmp_ ID returned by pages get.
# Save and read back a newly added component before adding a rule for it.
flowbase pages get pag_xxx > page-read.json
# Build page.json as {baseRevision: page-read.draftRevision, page: page-read.page},
# then remove the read-only Page keys listed by page-components.
# Or start from the complete offline example:
flowbase docs example pages.update --output page.json
flowbase pages update pag_xxx --json-file page.json

6.5 Records and attachments

records.list is a column projection of the current view (often only the title field). Use records.get for the full values object. values keys are field keys; select/lookup values are option/record IDs.

flowbase records list obj_xxx --page 1
flowbase records get obj_xxx rec_xxx
flowbase docs catalogs record-values
flowbase records create obj_xxx --json '{"values":{"name":"Ada","status":"opt_xxx","class":"rec_xxx"}}'
flowbase records upload-attachment obj_xxx fld_xxx ./contract.pdf
flowbase records update obj_xxx rec_xxx --json '{"values":{"attachments":["att_xxx"]}}'
flowbase records download-attachment obj_xxx rec_xxx att_xxx --output ./contract.pdf

6.6 Roles (production only)

export FLOWBASE_HOST=https://flowbase.deepflowagent.com
flowbase docs catalogs permissions      # function points + data scopes
flowbase roles options                  # apps and objects available for grants
flowbase roles create --json-file role.json
flowbase roles update rol_xxx --json-file role.json

7. Output and errors

  • Success: JSON on stdout, exit code 0. Use --no-pretty for compact JSON or --raw for the raw body.
  • Failures (missing key, invalid JSON, HTTP 4xx/5xx, network errors, cancelled confirmation) go to stderr with a clear message and exit code 1.
  • flowbase --help lists every command; flowbase <group> --help and flowbase <group> <action> --help show flags and, for JSON commands, an inline body example.

Error matrix

Scenario HTTP / CLI behavior
Missing or invalid API key 401; stderr tells you to set FLOWBASE_API_KEY or --api-key
Environment not enabled 403; request is rejected before business logic
Missing draft revision / concurrent edit 409; re-read with get and retry with the new baseRevision
Invalid request field or business rule 400 / 422; detail explains the reason
Resource not found 404
Destructive command without --yes in automation CLI refuses before any network call

8. Verification and cleanup

  • flowbase organization employees --query Ada --page 1: discover real approval references; follow the returned pages. organization departments returns the full department tree.
  • flowbase records list obj_xxx --view-id viw_xxx: query a specific active view. Invalid IDs return an error instead of silently selecting another view.
  • flowbase pages evaluate pag_xxx --json '{"values":{"done":true}}': evaluate draft rules on editor-style values. Returns rule overrides keyed by field/component IDs; does not render a page, validate record values, execute buttons, or write records.
  • flowbase pages clear-default pag_xxx --base-revision 4 --yes: remove only this page's default assignment, including on a deleted object, then use pages delete with the newly returned revision.
  • docs catalogs page-buttons, workflow-branches, approval-rules, and dashboard-runtime describe complex configuration and runtime limits.

Transport requests time out after 60 seconds. A lost write response has an unknown outcome: read the resource and current revision before retrying. The CLI never automatically retries writes.

Recent feature authoring

  • flowbase docs catalogs field-types: includes template autonumber configuration and date/time/datetime v2 creation examples. Numbering rules and temporal precision are immutable after creation; legacy numbering and temporal v1 remain supported.
  • flowbase docs catalogs planning: tree/Gantt view snapshots, candidate queries and optimistic plan updates. Read a full view with views list before updating; preserve displayMode, parentFieldId and gantt.
  • flowbase docs catalogs record-bundles: initialize, atomically submit and recover a parent/detail submission. Reuse a submission ID only with the identical request body.
  • flowbase docs catalogs formula-authoring: formula/rollup sources, validation, saved-record trials and runtime draft preview. Supports reuse of calculated results through single relations and multi-level rollups (up to 10 dependency levels). Validation returns dependencyGraph, or valid=false with code/errorPath for invalid dependencies. Pass fieldId when validating an existing field. Derived values are read-only.
  • flowbase docs catalogs workflow-python: record creation examples and failed-task retry. workflows python-runtime OBJECT new discovers current object bindings; workflows executions includes Python attempts, logs and canRetry.

Use records list OBJECT --view-id VIEW --view-layout for tree/Gantt projections; the default remains a flat list. records update accepts updatedAt and planViewId in its JSON body for version-checked plan edits. Formula validation and custom-sample trials require metadata read; saved-record trials also require record read; workflow retry requires both metadata write and record write and explicit CLI confirmation (--yes in automation).

Single-select formulas (4.21)

Use flowbase formulas sources OBJECT to discover local and one-hop single-select paths and their options (id, label, lifecycle). Use the returned stable option ID in is_option(path, "OPTION_ID"); option_label(path) reads the current label. Inactive options can model historical values. Raw single-select values cannot be compared with strings or returned directly.

Save a request such as this as formula.json, replacing the path and option ID with the source discovery response:

{
  "constraints": {"expression": "is_option(direction, \"OPTION_ID\")", "resultType": "boolean"},
  "samples": {"direction": "OPTION_ID"},
  "trial": true
}

Run flowbase formulas validate OBJECT --json-file formula.json. For label output, use option_label(direction) with resultType: "text.singleLine". Samples are stable option IDs or null; use recordId instead to test an accessible saved record. Inspect valid, optionReferences and trial.status: HTTP success alone does not mean the formula or trial succeeded. Save valid configuration through the existing fields create / fields update commands (metadata write permission).

These commands require a backend with 4.21 support. Existing CLI installations that already provide formulas sources and formulas validate can send the new expressions; updated offline help requires the updated CLI package. Employee CLI consumers continue reading calculated values and states through record commands.

Conditional query rollups (4.23)

Read flowbase docs catalogs conditional-rollup, then discover stable IDs with flowbase formulas collections OBJECT. A query source uses mode: "query", objectId, and 1–50 conditions of {fieldId, operator, currentFieldId}. All matching conditions must hold; optional fixed filters with match: "all" | "any" further narrow the matched records. No relation field is required. The catalog includes both standalone rollup and formula collection examples.

Submit the example with flowbase formulas validate OBJECT --json-file query.json. Include fieldId when editing an existing field. For a trial, use trial: true and a saved recordId; custom samples cannot trial collections. Validation requires metadata read; saved-record trials additionally require record read. Check valid and trial.status, then save valid constraints with fields create/update.

Matching uses saved current-record values. Empty matching inputs return pending/matching_empty; new records return saved_required. Successful authorized trials include collection parameters and candidate/matched counts. Full candidate permissions are checked before fixed filters; restricted results expose no totals or counts. Limits and timeouts never produce partial totals. See the catalog for compatible types, operators and budgets.

Detail Page value assignments (4.22)

flowbase docs catalogs page-value-rules describes authoring and the edit sequence. Page snapshots support setFieldValue and setDefaultFieldValue; sources can traverse up to five single-record relations. pages evaluate continues to preview field states and component visibility only.

Create and detail Page actions accept an optional explicit conversion: mapOptions with mappings: [{sourceOptionId, targetOptionId}], copyOptions for the same multiple-select definition through a relation, truncateDate for lower date precision, or extractDate for datetime → date in Asia/Shanghai. Read the catalog for the complete contract. Mappings use stable identities, never option names. Partial mappings warn; an unmapped runtime value fails the entire action. Disabled rules may retain inactive targets but must retain valid identities. Ordinary record writes remain strict and do not implicitly execute these conversions.

flowbase records get OBJECT RECORD
flowbase records page-value-context OBJECT RECORD
flowbase docs example records.page-value-event --output event.json
# Fill event.json with the returned pageId/revision, current edit delta,
# changed field IDs, and the preceding event's sources (initially {}).
flowbase records page-value-event OBJECT RECORD --json-file event.json

Context and event calls require both record read and write scopes, and do not save edits. Inspect errors, apply returned values to your edit delta, and carry sources into the next event's previousSources. Values use field keys, changed fields use IDs, and decimal assignments are returned as lossless strings. Defaults skip nonempty targets. Source/condition changes trigger assignment; simply reading or saving does not.

Save the merged edit delta with records update, including pageId, valueRuleRevision (the context's revision), and updatedAt. Configuration changes or stale record versions reject the save; reload and reconcile on 409. Page rules and field permissions are rechecked without rerunning assignments. Whole-record edits use the existing bundle-save / configVersion contract. Production context uses published configuration, while test Hosts use effective configuration.

Official app store (4.14)

Read flowbase docs catalogs app-store. Reads and checks require metadata read; install, publish and listing changes require metadata write and confirmation (--yes in automation).

flowbase app-store context
flowbase app-store list --query tasks --page 1
flowbase app-store get PACKAGE
flowbase docs example app-store.install-check --output installation.json
# Set the reviewed version and desired app name.
flowbase app-store install-check PACKAGE --json-file installation.json
# Copy version/name into install.json; add the returned check and a fresh UUID request_id.
flowbase app-store install PACKAGE --json-file install.json --yes
flowbase app-store installation REQUEST_UUID

Installation requires a clean enabled environment and creates configuration drafts only. Review and release them separately. The check binds the current Host environment, source package, name, API key and authorization; it expires after 15 minutes. Changed inputs require another check. After a timeout, query the receipt before retrying; reuse a request ID only with the identical body. Inspect status/error, since HTTP success may report processing or failed. A failed attempt needs a new check and request ID.

For eligible official publishers on the production Host:

  1. app-store sources discovers committed source apps; production drafts must be resolved first.
  2. app-store source-check --json '{"app_id":"APP"}' returns a signed source snapshot.
  3. docs example app-store.publish-check --output publication.json provides all required bilingual product fields. Fill them and the returned source, package ID and version, then run app-store publish-check --json-file publication.json.
  4. Review product content and checked resources. Add check, a fresh UUID request_id, and preview_confirmed: true; run app-store publish --json-file publish.json --yes.
  5. Query app-store publication REQUEST_UUID or app-store publications. Use app-store listing PACKAGE --json '{"listed":false}' --yes to withdraw, or true to relist.

Store request bodies use snake_case. --yes does not bypass publishing eligibility, production-only rules, portability validation or rechecks. These added commands need an updated CLI installation and a backend with the corresponding public routes; the package version is bumped separately during release.

Reusable iframe configuration (4.11)

flowbase iframes list|get|create|update|delete manages environment-specific iframe drafts. Create/update use --json-file; fields are label, icon (default IconWorld), color (default #2563EB), and a complete HTTPS url without credentials. Update and delete require baseRevision from iframes get; deletion also requires confirmation (--yes for noninteractive use).

flowbase iframes create --json-file iframe.json
flowbase iframes get ifr_example
flowbase iframes update ifr_example --json-file iframe-update.json
flowbase apps add-iframes app_example ifr_example
flowbase apps set-iframe-group app_example ifr_example --group-id grp_example
flowbase apps set-iframe-group app_example ifr_example --ungrouped
flowbase apps reorder-members app_example --json '{"items":[{"type":"iframe","id":"ifr_example"}]}'
flowbase apps remove-iframe app_example ifr_example --yes
flowbase iframes delete ifr_example --json '{"baseRevision":12}' --yes

Removal only unlinks one app. Deletion unlinks all apps atomically. Changes follow the existing commit/release/rollback workflow; they do not fetch, proxy or modify the external website. API keys use metadata:read/metadata:write; browser roles use iframe.manage for shared resources and app.manage for app associations.

Custom object views (4.12)

Read flowbase docs catalogs custom-views and flowbase views list OBJECT first. Save the following as view.json, replacing baseRevision with the current draftRevision:

{
  "baseRevision": 4,
  "name": "Custom overview",
  "enabled": false,
  "defaultFieldIds": [],
  "displayMode": "custom",
  "custom": {
    "runtimeVersion": "cv1",
    "source": "import React from 'react'; export default function View(){return <div>Hello</div>}"
  }
}
flowbase views create obj_example --json-file view.json
flowbase views list obj_example
flowbase views update obj_example viw_example --json-file view.json
flowbase views set-enabled obj_example viw_example --base-revision 6 --enabled --yes

Use the revision returned by the preceding write, rather than the example numbers. Custom views start disabled; preview in the browser and enable when ready. Updates replace the writable snapshot: preserve custom.source, custom.runtimeVersion, displayMode, filters and other configuration from views list, excluding the response-only id and status. Built-in/custom view types cannot be converted. JSX source is limited to 512 KiB UTF-8; the runtime supports react, three and flowbase. CLI configuration does not execute browser previews.

Record detail embedded pages (4.13)

flowbase docs catalogs page-components documents recordEmbed, supported at the top level or inside tabs of a detail Page. Add this component to an existing region column (or a tab's components):

{
  "kind": "recordEmbed",
  "label": "External panel",
  "showTitle": true,
  "url": "https://example.com/panel",
  "height": 480,
  "allowRead": false,
  "allowedFieldIds": []
}

Read flowbase pages get PAGE, preserve the existing components and IDs, and remove the read-only Page keys listed in docs catalogs page-components. Wrap the editable snapshot as {"baseRevision":4,"page":{...}}, then use flowbase pages update PAGE --json-file page.json (or pages create --json-file page.json for a new detail Page). For creation, omit the Page ID and IDs of new regions/components so the server allocates them.

Height must be an integer from 240 to 1600. URLs must use external HTTPS and exclude credentials, identity/context parameters and templates. To permit SDK reads, set allowRead: true and select at least one unique active field ID in allowedFieldIds; current record/field permissions still apply. The browser SDK is /sdk/record-embed.v1.js. The CLI configures the component; rendering and SDK calls run in the browser. flowbase-user has no metadata management commands.

Metadata

Release files for flowbase-cli 0.1.20

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for flowbase-cli 0.1.20
File Size Uploaded Provenance Actions
flowbase_cli-0.1.20.tar.gz 91.7 kB Without provenance Details for flowbase_cli-0.1.20.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for flowbase-cli 0.1.20
File Interpreter ABI Platform Actions
flowbase_cli-0.1.20-py3-none-any.whl Python 3 none any Details for flowbase_cli-0.1.20-py3-none-any.whl

Total release size: 153.8 kB

Release files / flowbase_cli-0.1.20.tar.gz

Download URL flowbase_cli-0.1.20.tar.gz
Size 91.7 kB
Tags Source
SHA-256 checksum
How to use checksums
f3c7f9eb70f1f48fd29f7274ead54d74c1c6e069ad435eb548b6177adb6287c9
BLAKE2b-256 checksum
How to use checksums
0311db0059d2194e30218c37beeb3df7f82b5a6a9113e3d61e456bebef3223de
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.13.16

Release files / flowbase_cli-0.1.20-py3-none-any.whl

Download URL flowbase_cli-0.1.20-py3-none-any.whl
Size 62.2 kB
Tags Python 3
SHA-256 checksum
How to use checksums
d4a57827a7d1fe25b00835ac4b1810529c543eb0f3d1009b66cd773485b61ff2
BLAKE2b-256 checksum
How to use checksums
7c64cc083cabbd3a7ef0c7abb9b6b6140d06f4f06b7718b8e96c522c98db05ca
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.13.16

Security for flowbase-cli 0.1.20

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page