Your AI team, working in PromptLayer
PromptLayer is a prompt management and observability platform for versioning, evaluating, running, and analyzing LLM applications.
Start your companyWhat your team can do in PromptLayer
115 tools to work with, connected with an API key.
Append one or more rows to a PromptLayer Table sheet in the API key's workspace. Text-column values can be set immediately; non-text cells are created stale and need a separate recalculation.
Attach an external-system identifier to an existing PromptLayer folder. Supply the folder ID, the external system's source namespace, and its identifier for the folder.
Attach an external system identifier to an existing PromptLayer prompt template. Supply the template ID, external namespace (source), and external ID. Returns the attached mapping; a mapping assigned to another entity conflicts.
Attach an external-system identifier to an existing PromptLayer registered tool. Supply its tool ID, source namespace, and external identifier. Returns the attached mapping; does not execute the registered tool.
Attach an external system identifier to a PromptLayer skill collection by UUID. Returns the attached source and external ID mapping; conflicting mappings return an error.
Attach an external-system identifier to an existing PromptLayer workflow. Supply the workflow ID, the external system's source namespace, and its identifier for the workflow.
Cancel an active PromptLayer Table sheet recalculation operation. The Table, sheet, and operation must belong to the API key's workspace. A successful response can mean no active operation was found: inspect cancelled_cell_count and message rather than treating success alone as proof of cancellation.
Cancel queued or running scorecard work for a PromptLayer table sheet. Optionally restrict cancellation to row indices and scorecard step IDs. The table and sheet must belong to the API key's workspace. Inspect cancelled_count to distinguish cancellation from a no-op.
Close a PromptLayer trace, permanently preventing further span ingestion for that trace through spans-bulk or OTLP /v1/traces. An already closed trace returns a conflict. The provider documents this endpoint, but its public API returned route-not-found during live verification; closure availability is unverified.
Configure scoring for a PromptLayer Table sheet using columns, boolean or numeric scoring, or custom code. Updates configuration and reports whether recalculation is required; does not queue or execute score calculation. All resource IDs must belong to the API key's workspace.
Create or update the active scorecard for a PromptLayer table sheet. Set its name, evaluated columns, evaluation steps, and aggregation settings. This configures the scorecard; calculation is queued separately. Resource IDs must belong to the API key's workspace.
Create a PromptLayer folder at the workspace root or inside a parent folder, optionally attaching external identifiers. The name must be unique within its parent. Uses the API key's workspace unless workspace_id is supplied.
Create a release label for a specific version of a PromptLayer prompt template. Supply the template ID, version number, and label name. This labels stored content without executing an LLM.
Create a PromptLayer registered tool and its initial version in the API key's workspace using an OpenAI function-calling definition. Optionally set a folder, external identifiers, and a sandbox-executable implementation. Returns the created tool and version; does not test-execute the tool.
Create an environment variable scoped to a registered PromptLayer tool, identified by numeric ID or name. Supply a value or leave it empty to create a placeholder. An existing key on the same tool returns a conflict; use the update action to change it.
Create a new version of an existing PromptLayer registered tool by numeric ID or name using an OpenAI function-calling definition. Optionally configure a sandbox-executable body and commit message. Returns the new version; does not invoke or test-execute the registered tool.
Create a PromptLayer skill collection in the API key's workspace, optionally with initial text files, a destination folder, provider, and external IDs. Uses the JSON creation endpoint; ZIP uploads are not supported by this action.
Create a PromptLayer Table in the API key's workspace, optionally in a folder. By default creates Sheet 1 with a Column A text column and one empty row; set create_default_sheet to false to create a table without sheets.
Add a column to a PromptLayer Table sheet in the API key's workspace. Provide a title and column type, with type-specific config and dependency edges when needed. Non-text columns generate cells for all existing rows.
Create a sheet in a PromptLayer Table in the API key's workspace. Omit source for a blank sheet with Column A and one empty row (201), or import a base64 CSV/JSON file or request log history asynchronously (202). For imports, poll the returned operation_id using the table operation endpoint to track completion.
Queue recalculation of a PromptLayer Table sheet, optionally selecting columns, zero-based rows, and cell statuses. Text columns cannot be recalculated. Returns operation and execution IDs for asynchronous work, a zero cell count for no-op, or a confirmation token to resend when confirmation is required. IDs must belong to the API key's workspace. Computed columns may incur costs.
Create a named snapshot of a PromptLayer Table sheet, or restore an existing version while creating a new version. Provide name for a snapshot; provide source_version_id to restore, optionally with name. Restoring changes the sheet's current contents. All IDs must belong to the API key's workspace.
Create a PromptLayer workflow, or a new version of an existing workflow. Use name for a new workflow (omit to auto-generate), or workflow_id or workflow_name to version an existing workflow. Supply the complete node list with at least one output node; optionally attach conditional edges, release labels, external identifiers, and required input variables.
Create a workspace environment variable in the API key's workspace. Supply a unique key and optionally a value, or create an empty placeholder. Returns the variable ID and metadata, including only the value suffix.
Delete PromptLayer registry entities by ID and type. Folders must be empty unless cascade is true, which recursively deletes their contents. Returns the affected entity count in moved_count.
Detach an external identifier from a PromptLayer folder using its folder ID, source namespace, and external ID value. Removes only the external-ID mapping; the folder and its contents remain.
Detach an external identifier from a PromptLayer prompt template using its template ID, source namespace, and external ID value. Removes only the external-ID mapping; the prompt template remains.
Delete a PromptLayer prompt template release label by its label ID. Removes the label without deleting the prompt template or its versions. Use Get Prompt Template Labels to find the label ID. Returns no content.
Delete one tool-scoped environment variable from a PromptLayer registered tool identified by its ID or name. Requires the environment variable's integer ID. This removes the stored variable, not the registered tool or workspace-scoped variables.
Detach an external identifier from a PromptLayer registered tool using its tool ID, source namespace, and external ID value. Removes only the external-ID mapping; the registered tool remains.
Detach one external ID mapping from a PromptLayer skill collection by its UUID, source, and external ID. Removes the mapping without deleting the collection.
Soft delete a PromptLayer Table by UUID, removing it from active list responses. The Table must belong to the workspace associated with the API key.
Delete a column and its cells from a PromptLayer Table sheet. Creates a new sheet version and may mark downstream dependent cells as stale. The Table, sheet, and column must belong to the API key's workspace.
Soft delete a sheet from a PromptLayer Table, removing it from active sheet list responses. Both IDs must belong to the API key's workspace, and edit permission is required.
Soft-delete the active scorecard for a PromptLayer table sheet. The table and sheet must belong to the API key's workspace. This removes the scorecard, not the table or sheet.
Detach an external identifier from a PromptLayer workflow using its workflow ID, source namespace, and external ID value. Removes only the external-ID mapping; the workflow remains.
Delete a PromptLayer workspace environment variable by its integer ID. The variable must belong to the workspace associated with the API key.
Get the prompts, version numbers, and labels using a prompt-template snippet by name or numeric ID. Optionally check usage for a specific snippet version.
Retrieve a prompt template version, release label, messages, model metadata, snippets, and optionally provider-ready kwargs without executing it.
Retrieve a PromptLayer registered tool definition by ID or name at a selected version or release label without test execution.
Read aggregated PromptLayer request-log analytics, including costs, tokens, latency percentiles, time series, and usage breakdowns. Optionally narrow results with text search or recursive filters; no request logs are modified.
Retrieve one PromptLayer request log with its recorded input, output, timing, usage, metadata, tags, and scores.
Get autocomplete suggestions for a PromptLayer request-log search field. Optionally filter by a case-insensitive prefix or JSON-encoded filter group. Value fields require metadata_key; prompt_id only filters the prompt field.
Retrieve a PromptLayer skill collection and its files as JSON at a selected version or release label.
Retrieve one PromptLayer Table's metadata and per-sheet row counts by UUID.
Retrieve a single cell by UUID from a PromptLayer Table sheet. Returns its value, computation status, and available request metrics for prompt-template cells. All IDs must belong to the API key's workspace.
Get a single scorecard calculation for a PromptLayer table sheet by its calculation UUID. Returns execution status, available aggregate score and verdict, criterion summaries, and drift metadata. The table, sheet, and calculation must belong to the API key's workspace.
Get the row-level scorecard breakdown for a zero-based row index in a PromptLayer table sheet, including aggregate score, verdict, step results, and drift or staleness metadata. Defaults to the latest calculation; provide calculation_id to read a specific calculation.
Get metadata for one sheet in a PromptLayer Table, including its row count, display order, and current version. Both IDs must belong to the API key's workspace.
Poll the status of an asynchronous PromptLayer Table sheet import using its operation ID. Returns progress, row counts, and completion or failure details. The Table and operation must belong to the API key's workspace.
Poll a PromptLayer Table sheet recalculation operation by its operation ID. Returns progress, cell status counts, and completion or failure status. The Table, sheet, and operation must belong to the API key's workspace.
Retrieve the current score payload for a PromptLayer table sheet, including calculation status, scoring configuration, per-column results, and aggregate scores. Does not configure or recalculate scores. The table and sheet must belong to the API key's workspace.
Retrieve score-history points across saved versions of a PromptLayer table sheet, with optional history range, resolution, and maximum point count. Returns points and sampling metadata. The table and sheet must belong to the API key's workspace.
Get the active scorecard for a PromptLayer table sheet, including its evaluation steps, latest and display calculation metadata, summaries, and progress. Returns a null scorecard when none is configured. The table and sheet must belong to the API key's workspace.
Retrieve one saved PromptLayer Table sheet version, including its snapshot and recorded changes. Table, sheet, and version IDs must belong to the API key's workspace.
Retrieve a PromptLayer trace by its trace ID, including its spans and any associated request-log IDs.
Retrieve a PromptLayer workflow definition by ID or name at a specific version or release label without running it.
Retrieve results of an already-started PromptLayer workflow execution. Returns the main output by default, or each output node's value, status, and errors with return_all_outputs=true. In-progress executions are also returned; inspect node statuses and retrieve again later. Does not run a workflow.
Start an asynchronous request-history import into an existing PromptLayer Table sheet. Provide specific request_log_ids or a structured filter_group to select logs in the API key's workspace. Returns an operation_id and status_url to poll; acceptance does not mean the rows have finished importing.
Start an asynchronous CSV import into an existing PromptLayer Table sheet. Provide the CSV filename and base64-encoded content. A queued response does not mean the import is complete; poll the returned status_url. The Table and sheet must belong to the API key's workspace.
Ingest OpenTelemetry traces using OTLP JSON, optionally gzip-compressed. Returns partialSuccess rejection details when some spans are rejected. Set close_after to close the submitted traces and reject subsequent writes.
Persist a batch of OpenTelemetry-style observability spans and optional associated request logs in PromptLayer without invoking a model. Supply nanosecond span timestamps. Optionally close all traces in the batch to reject subsequent writes. Inspect rejected_span_ids, rejected_trace_ids and rejection_reason for dropped spans, even when success is true.
Browse or search PromptLayer folders and registry entities in the API-key workspace, optionally recursively and with metadata.
List all external identifier mappings attached to a PromptLayer folder by its folder ID. Returns each mapping's source namespace and external ID, or an empty list when the folder has no external IDs.
List all external system identifiers attached to a PromptLayer prompt template by its numeric ID. Returns each mapping's source namespace and external ID. This endpoint returns the complete list without pagination.
List all release labels for a prompt template by name or numeric ID, including each label's associated version ID and version number.
Search and page through PromptLayer prompt templates and snippets without executing a model.
List environment variables for a PromptLayer registered tool by ID or name. Returns names, masked value suffixes and placeholder status, not full values.
List all external identifier mappings attached to a registered PromptLayer tool by its tool ID. Returns each mapping's source namespace and external ID, or an empty list when none are attached. Does not execute the tool.
List and filter tools defined in PromptLayer's tool registry without executing them.
List all external system identifier mappings attached to a PromptLayer skill collection by UUID. Returns each mapping's source and external ID.
List and filter PromptLayer skill collections available in the API-key workspace.
List one page of user and optionally system columns for a PromptLayer Table sheet.
Read one page of rows and cells from a PromptLayer Table sheet, with optional schema and execution aggregates.
List one page of row-level scorecard summaries for a PromptLayer table sheet. Defaults to the latest calculation; optionally select a calculation or filter by aggregate verdict. Pass the returned next_cursor as cursor to continue.
List active recalculation operations and cell status counts for a PromptLayer Table sheet in the API key's workspace. Returns the sheet version, total and pending cell counts, and progress for active operations.
List one page of saved versions for a PromptLayer Table sheet in the API key's workspace. Returns version details and pagination metadata. To continue, pass pagination.next_cursor as cursor with the same sort and order.
List one page of sheets in a PromptLayer Table, optionally filtering by referenced prompt.
Search and page through PromptLayer Tables by folder, name, or referenced prompt.
List all external system identifiers attached to a PromptLayer workflow by its numeric ID. Returns each mapping's source namespace and external ID. This endpoint returns the complete list without pagination.
List all release labels for a PromptLayer workflow by ID or name, including the version ID and version number each label points to.
Search and page through PromptLayer workflows without running them.
List environment variables for the API key's PromptLayer workspace. Returns keys, scope IDs, timestamps, empty-value flags, and only the last four characters of stored values, rather than plaintext values.
Log an existing model request and response in PromptLayer with timing, usage, cost, tags, metadata, and optional prompt association. This records supplied data without invoking an LLM. Chat message content must be an array of content blocks, not a plain string. Returns the created request log ID.
Convert a PromptLayer Table sheet's existing legacy score configuration into scorecard criteria. Optionally delete the legacy score configuration after conversion. Returns the scorecard, converted count, skipped scores, and sheet version. Resource IDs must belong to the API key's workspace.
Move PromptLayer registry entities into a destination folder or to the workspace root. Supply each entity's ID and type; omit folder_id or set it to null to move to the root. Returns the number of moved entities.
Move an existing PromptLayer release label to a different version of the same prompt template. Supply the label ID and destination version number; the label's name and identity are preserved.
Create a new version of an existing PromptLayer template by patching its latest version, a specific version, or a release label. Patch chat messages/tools/functions or completion content by index or replace whole arrays, tune model parameters, and optionally move release labels. Omitted fields are preserved; explicit null removes supported settings. This stores a new version without executing an LLM or modifying the base version.
Partially update a PromptLayer workflow by ID or name, creating a new version from the latest or a specified base version. Add, deep-merge, or remove nodes; replace dependencies, inputs, or edges; and attach release labels. The resulting workflow must retain an output node. Does not execute the workflow.
Publish a PromptLayer registry prompt template and a new version of its chat or completion content. Creates a template or adds a version to the existing prompt name. Optionally attach tags, model metadata, release labels, and external IDs. This stores the prompt without executing an LLM.
Read custom aggregate analytics over PromptLayer request logs. Compute counts, numeric aggregates, percentiles, grouped or time-series charts, and multi-series ratio insights, optionally narrowed with search and recursive filters. Does not modify request logs or save charts.
Read custom aggregate analytics over PromptLayer traces and spans with recursive filters. Compute counts, costs, tokens, durations, percentiles, grouped charts, histograms, or multi-series results. Each chart must use exclusively trace-level or span-level fields; span charts cannot use timeSeries. No traces or saved charts are modified.
Trigger recalculation for one PromptLayer Table cell. Returns an execution ID when recalculation starts, or a message when no cells need recalculation. All IDs must belong to the API key's workspace. Computed columns may incur execution costs depending on their type.
Trigger recalculation for a batch of PromptLayer Table cells by ID. Returns an execution ID when recalculation starts, or a message when no cells need recalculation. All IDs must belong to the API key's workspace. Computed columns may incur execution costs depending on their type.
Queue recalculation of an existing PromptLayer Table sheet score configuration. Scoring must already be configured; this action does not configure scoring or wait for calculation to finish. The Table and sheet must belong to the API key's workspace.
Queue recalculation of an existing PromptLayer table sheet scorecard, optionally limited to row indices and step UUIDs. The scorecard must already be configured. Returns a calculation ID and status without waiting for completion. Resources must belong to the API key's workspace.
Retrieve a prompt template by name or ID and render provider-ready llm_kwargs with input_variables using POST. Optionally select a version or release label and obtain provider-specific llm_kwargs with model parameter overrides. Does not execute an LLM call or save a template version. Use GET_PROMPT_TEMPLATE for the raw template read without input-variable substitution. The prompt_template response may retain raw placeholders; the substituted content is in llm_kwargs.
Resolve a PromptLayer folder's ID from its dot-separated folder path within a specified workspace. Returns the folder ID for subsequent folder operations; a nonexistent path returns a not-found error.
Start an execution of a PromptLayer workflow by name, using its latest version or one specified label or version number. Returns an execution ID, not the completed output; retrieve results with Get Workflow Execution Results. Optionally supply inputs, metadata, or an HTTP callback URL for asynchronous delivery.
Create a new saved version of a PromptLayer skill collection by UUID, name, or root path. Send file text updates, path moves, and deletions as JSON, with an optional commit message and release label.
Search PromptLayer request logs by text or structured filters and return one rate-limited page sorted by operational fields.
Execute a PromptLayer registered tool's code in its sandbox by ID or name, optionally selecting a version or label and testing unsaved code or a tool definition. Executes real code and may cause its configured side effects. User-code failures are returned as result.status='error' in a successful HTTP response.
Associate an existing PromptLayer request log with an existing group using their integer IDs. Returns confirmation of the association.
Associate custom metadata with an existing PromptLayer request log using its request ID. Returns a confirmation message.
Associate an existing PromptLayer request log with a named prompt template, optionally specifying its version or release label and the variables used to render it. This updates the log's prompt association; it does not run a model.
Record a score from 0 to 100 on an existing PromptLayer request log. Supply the request_id returned by logging a request and optionally a score name; omitting the name records the default score.
Rename an existing PromptLayer folder by ID. The new name must be unique within its parent folder. Returns the updated folder metadata.
Replace the stored value of an existing PromptLayer registered tool environment variable by tool ID or name and variable ID. Requires a non-empty value and leaves the variable name unchanged. Returns metadata and the value's last four characters, not the full stored value.
Rename a PromptLayer skill collection identified by UUID, name, or root path.
Update a PromptLayer Table's title or folder by UUID. The table and folder must belong to the API key's workspace. Omitted fields remain unchanged.
Edit the display value or structured value of a PromptLayer table cell. Only text-column cells can be edited directly; computed cells recalculate automatically. All IDs must belong to the API key's workspace.
Update a PromptLayer Table sheet column's title, configuration, or dependencies. Returns whether existing cell values require recalculation and affected column IDs. All IDs must belong to the API key's workspace.
Update a PromptLayer Table sheet's title or zero-based display position. The Table and sheet must belong to the API key's workspace.
Replace the stored value of an existing PromptLayer workspace environment variable by ID. Requires a non-empty value; returns metadata and the value's last four characters, not the full value. The variable's name is unchanged.
Create or update a PromptLayer registry prompt template identified by an external source and ID. Creates the template when the mapping does not exist; otherwise publishes a new version of the same template. Supply registry metadata and chat or completion content, with optional release labels and additional external IDs. Stores the prompt without executing an LLM.
Nothing matches that.
More in Artificial intelligence
See them all βPerplexity AI provides conversational AI models for generating human-like text responses
28 toolsβ MMem0 is an universal, self-improving memory layer for LLM applications.
47 toolsβ EEntelligence leverages artificial intelligence to provide insights, recommendations, and predictive analytics for businesses seeking data-drivenβ¦
2 toolsβ RRetellAI captures calls and transcripts, enabling businesses to analyze conversations, extract insights, and enhance customer interactions in oneβ¦
70 toolsβ MMetatext AI specializes in natural language processing and text generation, helping organizations automate writing tasks, sentiment analysis, orβ¦
13 toolsβ DDataRobot is a machine learning platform that automates model building, deployment, and monitoring, enabling organizations to derive predictiveβ¦
836 toolsβ