--- title: "Roll a flow back to an earlier version" method: POST path: "/external/v1/flows/{flow_id}/versions/{number}/restore" tags: ["Flows"] --- # Roll a flow back to an earlier version `POST /external/v1/flows/{flow_id}/versions/{number}/restore` ## Request - Base URL: `https://api.databar.ai` (the document declares no server; this is the origin it was published from) - URL: `https://api.databar.ai/external/v1/flows/{flow_id}/versions/{number}/restore` - Auth: API key in header `x-apikey` Make an older version current. Nothing is overwritten: the old config is written as a new version, so the rollback is itself revertible. A version that no longer validates — its enrichment was deleted, its dataset changed — is refused with the reasons. ## Path parameters - `flow_id` string, required - `number` integer, required ## Request body - RestoreFlowVersionIn - `internal_version` string, nullable: Optional concurrency check, same semantics as on update. ## Response `200` OK - RestoreFlowVersionOut - `flow` FlowDetailOut, required: ``FlowOut`` plus the pieces an editing client needs. Without ``config`` and ``internal_version`` a caller cannot do a read-modify-write at all, which is why the API could only run flows. - `id` string, required - `name` string, required - `description` string, required - `inputs` FlowInputOut[], required - `id` string, required - `description` string, required - `type` string, required - `required` boolean, required - `outputs` FlowOutputOut[], required - `id` string, required - `key` string, nullable - `name` string, nullable - `response_field_id` integer, nullable - `created_at` string, required - `updated_at` string, required - `config` FlowDefinition, required - `inputs` FlowDefFunctionInput[] - `id` string, required - `description` string - `type` string: Same vocabulary as RequestParam.type_data. - `required` boolean: If true, input_bindings must include this id when attaching the function to a table. - `nodes` union[], required - union - FlowDefEnrichmentNode - `id` string, required - `type` 'enrichment' - `label` string - `on_error` 'stop' | 'skip': What to do when the enrichment call fails: stop the whole run or skip this node. - `on_no_data` 'stop' | 'skip': What to do when the enrichment returns no data: stop the run or skip this node. - `data` FlowDefEnrichmentNodeData, required - … - FlowDefConditionNode - `id` string, required - `type` 'condition' - `label` string - `on_error` 'stop' | 'skip': What to do when formula evaluation raises an exception. - `data` FlowDefConditionNodeData, required - … - FlowDefJqNode: Transform a JSON value using a jq expression. - `id` string, required - `type` 'jq' - `label` string - `on_error` 'stop' | 'skip': What to do when the jq program raises: stop the run or skip this node. - `data` FlowDefJqNodeData, required: Data for a JQ transform node — runs a jq expression on one or more JSON inputs. - … - FlowDefMergeNode: Concatenate multiple string inputs into a single value. - `id` string, required - `type` 'merge' - `label` string - `on_error` 'stop' | 'skip' - `data` FlowDefMergeNodeData, required: Data for a merge/concatenate node. - … - FlowDefMathFunctionNode: Evaluate an arithmetic expression with SymPy (Math Function tool). - `id` string, required - `type` 'math_function' - `label` string - `on_error` 'stop' | 'skip' - `data` FlowDefMathFunctionNodeData, required: Data for a Math Function node — evaluates a SymPy arithmetic expression. - … - FlowDefTableLookupNode: Query another Databar table and return matching rows as JSON. - `id` string, required - `type` 'table_lookup' - `label` string - `on_error` 'stop' | 'skip' - `data` FlowDefTableLookupNodeData, required: Data for a table-lookup node — queries another Databar table. - … - FlowDefWaterfallNode: Run a multi-provider waterfall enrichment pipeline. - `id` string, required - `type` 'waterfall' - `label` string - `on_error` 'stop' | 'skip': What to do when the waterfall fails: stop the run or skip this node. - `on_no_data` 'stop' | 'skip': What to do when no provider returns data: stop the run or skip this node. - `data` FlowDefWaterfallNodeData, required: Data for a waterfall node — runs ordered enrichments until one succeeds. - … - `named_outputs` FlowDefNamedFlowOutput[], nullable - `id` string, required - `key` string, required - `name` string, required - `source` FlowDefFlowOutputSource - `node_id` string, required - `ref` string, nullable - `mapping` FlowDefFlowOutput: Flow-level output mapping: reshape the run result into a custom shape. Two mutually-exclusive modes: * :attr:`fields` — structured mapping. Each entry is a :data:`NodeInput` whose ``id`` becomes the key in the returned ``outputs`` object. * :attr:`template` — a raw JSON string the user writes directly, with ``{{ input.x }}`` / ``{{ nodes.id.port }}`` tokens resolved at runtime. Supports arbitrary nested JSON, not just a flat object. When ``FlowDefinition.output`` is set, the executor returns the reshaped result instead of the raw namespaced ``{node_id: {port: value}}`` dict. - `fields` union[]: Output field bindings; each `id` is the key in the returned outputs object. - union - … - `template` string, nullable: Raw JSON template with `{{ input.x }}` / `{{ nodes.id.port }}` tokens. Overrides `fields`. - `combine` string[] - `combine_fields` FlowDefCombinedOutputField[] - `id` string, required - `output_id` string, required - `path` string - `output` FlowDefFlowOutput: Flow-level output mapping: reshape the run result into a custom shape. Two mutually-exclusive modes: * :attr:`fields` — structured mapping. Each entry is a :data:`NodeInput` whose ``id`` becomes the key in the returned ``outputs`` object. * :attr:`template` — a raw JSON string the user writes directly, with ``{{ input.x }}`` / ``{{ nodes.id.port }}`` tokens resolved at runtime. Supports arbitrary nested JSON, not just a flat object. When ``FlowDefinition.output`` is set, the executor returns the reshaped result instead of the raw namespaced ``{node_id: {port: value}}`` dict. - `fields` union[]: Output field bindings; each `id` is the key in the returned outputs object. - union - FlowDefInputRef: Binds a node input slot to a function-level input. - … - FlowDefNodeRef: Binds a node input slot to an output of another node. - … - FlowDefStaticRef: Binds a node input slot to a hardcoded value. - … - FlowDefStaticWithRefsInput: Binds a node input slot to a template string with ``{ref_id}`` substitutions. Used for prompt-text style fields where the resolved value is a literal with references interpolated inline, e.g. ``"Tell me about {company}"`` where ``company`` is a nested ref to a function input or upstream node output. Tokens supported: * ``{name}`` — required ref; raises ``ValueError`` if no matching entry in :attr:`refs` resolves to a value (i.e. the ref is missing or the upstream port is unset). An unmatched name that equals a function input id is treated as that input (MCP/UI often persist ``{domain}`` with empty refs). * ``{?name}`` — optional ref; falls back to ``""`` when missing. The token character set matches the table-side rich editor (``[a-zA-Z0-9_.]``) so prompts copied from a table workflow keep working. - … - `template` string, nullable: Raw JSON template with `{{ input.x }}` / `{{ nodes.id.port }}` tokens. Overrides `fields`. - `ui` object, required - `internal_version` string, required - `version` FlowVersionOut, required - `number` integer, required - `name` string, required - `description` string, required - `changed_fields` string[], required - `source` string, required: Where the edit came from: ui, api, mcp, admin or system. - `created_at` string, required - `created_by` string, nullable, required - `restored_from` integer, nullable: Version this one was restored from, if it came from a rollback. --- [API](https://skmtc.dev/databar/apis/external-api.md) - [All operations](https://skmtc.dev/databar/apis/external-api/llms.txt) - [OpenAPI document](https://skmtc.dev/databar/apis/external-api/revisions/efe15f8bb9bb?raw)