> ## Documentation Index
> Fetch the complete documentation index at: https://docs.edges.run/llms.txt
> Use this file to discover all available pages before exploring further.
# Product Updates
> New updates and improvements!
### Improvements
* **Extract LinkedIn People** (`linkedin-extract-people`) — Experience `date` fields now return employment date ranges only. `profile_country` is inferred from recent experience locations when LinkedIn does not expose it directly (remote/hybrid locations excluded). [Extract LinkedIn People](/v1/api/actions/linkedin-extract-people)
* **Extract LinkedIn People Experiences** (`linkedin-extract-people-experiences`) — Fixed `title` and `company_name` mapping; tenure values are no longer placed in `company_name`. [Extract LinkedIn People Experiences](/v1/api/actions/linkedin-extract-people-experiences)
* **LinkedIn Search People** (`linkedin-search-people`) — Async runs that hit a monthly LinkedIn limit (`LIMIT_REACHED`) are now automatically retried when the limit resets instead of failing permanently. [Search LinkedIn People Async](/v1/api/actions/linkedin-search-people-async)
* **Async & schedule runs** — Maximum automatic retries increased from 5 to 10 (up to 11 attempts per run including the initial attempt). [Execution Modes](/v1/runs/overview#async-mode)
### API
* **`NO_DATA_AVAILABLE`** — New error label returned when no data is found for an input (`No data was found for this input.`). [Error Reference](/v1/error-reference)
* **Extract Sales Navigator Company** (`salesnavigator-extract-company`) — Output field `profile_image_url` is deprecated. [Extract Sales Navigator Company](/v1/api/actions/salesnavigator-extract-company)
### Documentation
* **Identity rate limits** — Updated default identity-wide caps: 50/min, 500/15m, 2,500/4h, 15,000/24h. [LinkedIn Smart Limits](/v1/linkedin/limits)
* **Rate limits** — Documented monthly `LIMIT_REACHED` auto-retry for async people search. [Rate Limits](/v1/runs/rate-limits)
* **Extract LinkedIn People** — Documented `profile_country` inference and experience `date` field behavior. [Extract LinkedIn People](/v1/api/actions/linkedin-extract-people)
* **Extract LinkedIn People Experiences** — Documented `title` / `company_name` field mapping. [Extract LinkedIn People Experiences](/v1/api/actions/linkedin-extract-people-experiences)
### API
* **Extract LinkedIn People** (`linkedin-extract-people`) — Profile enrichments quota restored to full capacity: **10,000 / 24 h** (`LK_PROFILE_ENRICHMENTS`). Most previously unavailable fields are back; three output fields remain missing or `null`: `birth_date`, `linkedin_thread_id`, and `connected_at`. [Extract LinkedIn People](/v1/api/actions/linkedin-extract-people)
### Documentation
* **Extract LinkedIn People** — Updated unavailable-fields notice and limits reference to reflect restored capacity and the three remaining gaps. [LinkedIn Smart Limits](/v1/linkedin/limits)
### API
* **Mark LinkedIn Message as Read** — New action `linkedin-mark-message-as-read` (live/async/schedule). Mark a conversation as read using its `linkedin_thread_id`. Requires an [Engagement Identity](/v1/identities/engagement). [Mark LinkedIn Message as Read](/v1/api/actions/linkedin-mark-message-as-read)
### API
* **Extract LinkedIn Page Followers** (`linkedin-extract-page-followers`) — Incremental sync (`sync_mode: "incremental"`) is now supported in async and schedule modes. Fetch only new followers since the last run. Requires an [Engagement Identity](/v1/identities/engagement). [Extract LinkedIn Page Followers](/v1/api/actions/linkedin-extract-page-followers)
### API
* **Sales Navigator InMail Profile** — Improved error handling when a target user blocks us: we now check remaining InMail credits and return proper errors.
* **Sales Navigator ID ↔ LinkedIn handle mapping** — New actions [`linkedin-extract-salesnavigator-profile-id`](/v1/api/actions/linkedin-extract-salesnavigator-profile-id) (handle/URL → SN ID) and [`linkedin-extract-linkedin-profile-handle`](/v1/api/actions/linkedin-extract-linkedin-profile-handle) (SN ID/URL → handle).
### Improvements
* **Extract LinkedIn People** (`linkedin-extract-people`) — Added `number_share_connections` to the output for visibility into mutual connections. [Extract LinkedIn People](/v1/api/actions/linkedin-extract-people)
* **Extract LinkedIn People Experiences** (`linkedin-extract-people-experiences`) — Improved handling of `job_time_period` values that do not include a month. [Extract LinkedIn People Experiences](/v1/api/actions/linkedin-extract-people-experiences)
* **Extract LinkedIn Event Attendees** (`linkedin-extract-event-attendees`) — Fixed data handling for private/blocked profiles with consistent nullability and safe access. The number of results returned may differ from the count shown on LinkedIn when some attendees have private profile settings. [Extract LinkedIn Event Attendees](/v1/api/actions/linkedin-extract-event-attendees)
### Documentation
* **Only extract unique profile** — Clarified the `only_extract_unique_profile` behavior in docs.
* **Sales Navigator schema differences** — Documented fields temporarily unavailable when data is sourced from Sales Navigator on `linkedin-extract-people` and `linkedin-extract-company`.
* **New identity ramp-up** — Documented 5-day and first-day hourly capacity ramp-up for newly connected identities. [LinkedIn Smart Limits](/v1/linkedin/limits#new-identity-ramp-up)
### Platform
* **LinkedIn cookies lifecycle** — Persist additional cookies on refresh and login to improve session stability and stealth.
* **LinkedIn Search Generic** — Improved `linkedin_profile_id` handling and validation to reduce malformed or ambiguous IDs.
### Documentation
* **LinkedIn Smart Limits** — Quotas and per-minute caps in the limits reference are aligned with `IntAccLimits` (24 h and \~31 d windows). Added default rate caps when an action omits them (5/s, 30/min, 300/15 min, 1,200/4 h, 7,000/24 h), plus limits for profile components, saved searches, posts, activity extracts, and InMail. [LinkedIn Smart Limits](/v1/linkedin/limits)
### API
* **Extract Sales Navigator Messages** — New action `salesnavigator-extract-messages` (live/async/schedule). [Extract Sales Navigator Messages](/v1/api/actions/salesnavigator-extract-messages)
* **Extract Sales Navigator Conversations** — New action `salesnavigator-extract-conversations` (live/async/schedule). [Extract Sales Navigator InMail Conversations](/v1/api/actions/salesnavigator-extract-conversations)
### Correction
* **Uniform `linkedin_profile_url` format** — `linkedin_profile_url` values no longer include `miniProfileUrn` query parameters, so the same canonical URL shape is used consistently across Edges endpoints.
* **Current company in header** — On [`linkedin-extract-people`](/v1/api/actions/linkedin-extract-people), `company_name` now matches the company name shown in the profile **header** when the **`highlights`** parameter is enabled.
### Improvements & Fixes
* **LinkedIn protections** — Stricter **per-action** limits, broader **async** safeguards, and **higher minimum delays** on sensitive extractors to cut throttling risk and better protect identities.
* **Extract LinkedIn People** (`linkedin-extract-people`) — Fixed profile **country** resolution. [Extract LinkedIn People](/v1/api/actions/linkedin-extract-people)
* **Extract LinkedIn Company Viewers** — Fixed **missing profile pictures** in viewer results. [Extract LinkedIn Company Viewers](/v1/api/actions/linkedin-extract-company-viewers)
### API
* **Search Sales Navigator Company Metrics** — New action `salesnavigator-search-company-metrics`: total-count metrics for a Sales Navigator **company** search, with its own request/response schema.
* **Search Sales Navigator People Metrics (rename)** — `salesnavigator-search-metrics` is renamed to `salesnavigator-search-people-metrics`.
* **Backward compatibility** — `salesnavigator-search-metrics` remains accepted as an alias; existing callers are not broken.
### App
* **Search identities by UID** — You can search for identities by their **UID** in the [Identities](https://app.edges.run/identities) page in the dashboard.
### Improvements
* **LinkedIn outreach (async only)** — `linkedin-inmail-profile`, `linkedin-message-profile`, `linkedin-visit-profile`, `linkedin-connect-profile`, and `salesnavigator-inmail-profile` now have stricter pacing in **async** : big volume will take a little bit longer as usual; **live mode is unchanged** : you still have to space your call yourself.
* **LinkedIn checkpoint / CAPTCHA** — Fewer automatic attempts to solve captcha during identity login checkpoints (reduces **524** timeouts). If resolution still does not complete in time, responses may return a **`CAPTCHA`** checkpoint; poll [`POST /v1/identities/{identity_uid}/integrations/linkedin/checkpoint`](/v1/api/linkedin/checkpoint) with `{}` until it clears, then handle any next checkpoint per [Resolve LinkedIn Checkpoint](/v1/api/linkedin/checkpoint).
### Documentation
* **Resolve LinkedIn Checkpoint** — Documented **CAPTCHA after maximum attempts**, reduced captcha attempt limit (**524** mitigation), polling with `{}`, and follow-on checkpoints. [Checkpoint API](/v1/api/linkedin/checkpoint)
### API
* **Incremental Sync Mode** — New `sync_mode` parameter (`full` or `incremental`) for async and scheduled runs. Fetch only new data since the last retrieval instead of the full dataset. Available on engagement-free actions with Engagement Identities only. [Learn more](/v1/runs/overview#incremental-sync-mode)
* **Continue a Run** — New endpoint [`POST /v1/runs/{run_uid}/continue`](/v1/api/runs/continue) to trigger incremental updates on async runs.
* **Cursor-based pagination on run outputs & inputs** — `/v1/runs/{run_uid}/outputs` and `/v1/runs/{run_uid}/inputs` now use cursor-based pagination by default via `X-Pagination-Next`. The `offset` parameter is deprecated. [Pagination guide](/v1/runs/pagination#paginating-run-outputs--inputs)
### Improvements
* **Proxy Retry Logic** — `502` proxy errors are now retried internally with sticky IP fallback before being returned, reducing transient failures.
* **Identity Sync** — Webhook notifications are no longer sent during pending in-app challenges.
* `LK_BAD_COOKIE` errors now return the affected `identity_uid`.
* **Identity `is_active` property** — Identities now expose an `is_active` field. Inactive engagement identities are excluded from subscription renewals.
* **Output: `headline` deprecated, use `summary`** — On Search Sales Navigator People, Saved People, Company Employees, Search LinkedIn People, and Search LinkedIn Company Employees, the `headline` output is deprecated (retroactive). Use **`summary`**, which now correctly contains the profile summary; `headline` was a naming mistake. [Action docs](/v1/api/actions/salesnavigator-search-people).
* **Extract Sales Navigator Company** — New action. Inputs: **`sales_navigator_company_url`** or **`sales_navigator_company_id`**. Outputs include **`country`**, **`city`**, **`postalCode`** and full company profile data. [Action docs](/v1/api/actions/salesnavigator-extract-company).
### App
* **Subscription & pricing display** — Credits plan: discounted **base\_prices** and **total\_price** when a credit discount applies; **total\_price** when a global discount applies. **Engagement identities**: count and **total\_price** (with discount when applicable) are shown; engagement price is added to the selected plan total so you see the full amount (credits + engagement identities). Subscription modal shows the credit plan total with or without discounts applied.
* **`LIMIT_REACHED` errors** — Now include **`params.appendix`** and **`params.timespan`** to distinguish the limit: third-party API (`API_FUSE_LIMIT` daily, `API_RATE_LIMIT` per minute) vs product limits (`MINUTE_LIMIT`, `DAILY_LIMIT`). Integrations should inspect **`appendix`** for better retry logic. [Rate Limits](/v1/runs/rate-limits#differentiating-limit_reached-with-appendix) · [Error Reference](/v1/error-reference).
### Documentation
* **Engagement Identity pricing** — Billing is based on the **peak number of active Engagement Identities** during the month. Documented when an Engagement Identity is active (LinkedIn integration added or action called) vs inactive (30 days without any action, or 5 days after syncing without calling an action). [Engagement Identities](/v1/identities/engagement#pricing--billing)
* **Error Reference** — Error page updated: expanded [Quick Reference Table](/v1/error-reference#quick-reference-table) with all error labels, HTTP codes, categories, and retry guidance; [Error scopes](/v1/error-reference#understanding-error-scopes); [Handling 429](/v1/error-reference#handling-rate-limit-errors-429) (Edges vs LinkedIn, `params.appendix` / `params.timespan`); [Handling 424](/v1/error-reference#handling-linkedin-errors-424) LinkedIn errors with example code. [Error Reference](/v1/error-reference)
### API
* **Extract Sales Navigator Company** — New action to extract full Sales Navigator company data (revenue, description, employee counts, etc.) from a company page. [Extract Sales Navigator Company](/v1/api/actions/salesnavigator-extract-company)
### Improvements & Fixes
* **Extract LinkedIn Conversations** — Maximum conversations limit increased to 15,000; related validations updated. Error handling improved.
* **Extract LinkedIn School Alumnis** — Maximum alumni results limit set to 1,000.
* **Extract LinkedIn Company Affiliates** — Added safe optional chaining when accessing organization data to avoid runtime errors on missing fields.
### API
* **Extract LinkedIn Company Viewers** — New action to extract viewers of a LinkedIn company page. Requires page admin access and a Premium page; returns viewers from the past year. [Extract LinkedIn Company Viewers](/v1/api/actions/linkedin-extract-company-viewers)
* **Extract LinkedIn People Educations** — New action to extract education history (schools, degrees, fields of study, dates). [Extract LinkedIn People Educations](/v1/api/actions/linkedin-extract-people-educations)
* **Extract LinkedIn People — Education output** — New fields: `title` (degree + field of study), `field_of_study`, `school_description`. **Breaking:** `degree_name` is now correctly filled (first part of `title` before comma) or empty; it was previously sometimes filled with `field_of_study`.
* **ID type changes** — Several IDs changed from string to integer: `linkedin_school_id`, `linkedin_company_id`, `original_linkedin_company_id`, `company_id` in extract-people, experiences, company, company-employees-insights, recruiter-lite-search-people.
* **LinkedIn experience outputs** — Added `job_contract_type`, unblocking richer profile refresh data for migrations.
* **Outputs APIs** — Added `extracted_at` so consumers can reliably track when data was fetched.
* **Extract LinkedIn People Post/React/Comment Activity** — Output restructure:
* **Deleted:** `linkedin_original_post_url`
* **Deprecated:** `original_author`
* **Added** `source_post`: source post info (2nd level) with `linkedin_post_id`, `linkedin_post_url`, `content`, `author`; returns `null` when no 2nd level post
* **Added** `linkedin_activity_id`: unique ID depending on post type (original, simple reshare, reshare with comment)
* **Added** `activity_author`: person who performs the activity
* **Updated** `author`: now correctly filled with the author of the first-level (visible) post
* [Post levels table](https://www.notion.so/linkedin-extract-people-post-activity-2e818c28ccb280a5981dc18a54682e17?pvs=21)
### Improvements & Fixes
* **Billing** — Next billed amount displayed on the platform now takes into account any discount applied on the subscription.
* **Sales Navigator** — Fixed company keyword search when spaces were double-encoded, restoring results for multi-word queries.
* **Runs** — Treated `STATUS_404` and `NO_RESULT` as successful “no data” outcomes, reducing false failed runs while keeping transparency via `last_info`.
* **Experience location** — Improved completeness by inheriting the parent location when sub-experiences omit it.
* **Accounts** — Prevents false "no access" errors for some valid accounts.
### Post Activity — Known Limitations
* **Group posts:** When the post is a group post, `author` may be empty (not handled).
* **Component posts:** Some posts (e.g. `article`, `conversationComponent`) may not have an `author` or require alternative resolution.
### API
* **Get Run Inputs** — New endpoint to list all inputs for a run (including failed) and their errors. [Get Run Inputs](/v1/api/runs/inputs)
* **Runs** — Error responses 404/409/422 added on run endpoints; `auth_type` removed from account schemas.
### Documentation
* **Callbacks** — Data completeness: “all” vs “final” modes, how to verify and fetch results. [Managing Callbacks](/v1/runs/callbacks)
* **Extract Post Activity** — “Understanding Author Fields” (author, activity\_author, source\_post). [Extract LinkedIn People Post Activity](/v1/api/actions/linkedin-extract-people-post-activity#understanding-author-fields)
* **Run Outputs** — Not available for live runs (explicit 422). [Get Run Outputs](/v1/api/runs/outputs)
* **Runs overview** — Get run inputs added to [endpoints table](/v1/runs/overview#available-endpoints-for-runs).
* **Search Company Employees** — Reuse a people search URL on a company in one call. [Search Sales Navigator Company Employees](/v1/api/actions/salesnavigator-search-company-employees)
* **Message & InMail** — Voice/video format detection and conversion; Sales Navigator InMail unchanged. [Message](/v1/api/actions/linkedin-message-profile) · [InMail](/v1/api/actions/linkedin-inmail-profile)
### Improvements & Fixes
* **Run Outputs** — Pagination fix; 422 for live runs instead of empty response.
* **Workspace** — OAuth page white-label (your logo); credit override no longer blocks when below plan.
* **Retries** — Internal automatic retries (proxy, limits, Sales Navigator). Do not change your implementation, but decrease error rates.
* **Actions** — `linkedin_profile_id` in more outputs; Search Metrics returns 0 values and supports `#` in URLs; Post Activity returns `linkedin_activity_id`; identities list pagination fixed.
### What This Means for You
* **Recover missed data** — New inputs endpoint and clearer callback rules.
* **Voice & video** — Message and InMail send audio/video as voice or video messages where supported.
* **Company search** — Reuse a people search on a company in one call.
* **Fewer manual retries** — More cases trigger automatic retry.
* **Clearer API** — Pagination fixed; live runs return explicit error.
* **Workspace** — OAuth logo, credit override fix.
### Deprecations
#### Page-Based Pagination Deprecated (Live Mode)
The `page` parameter is now **deprecated** in favor of cursor-based pagination on 38 live mode endpoints.
| Parameter | Status |
| --------- | ---------------------------------------------------- |
| `page` | **Deprecated** (still functional, removed from docs) |
| `token` | **Deprecated** (still functional, removed from docs) |
| `cursor` | **New** - recommended approach |
> The deprecated parameters remain backward compatible. Existing integrations will continue to work, but we recommend migrating to cursors for new implementations.
**Migration steps:**
1. Replace `page` parameter with `cursor`
2. Replace `token` parameter usage with `cursor`
3. Use the `cursor` value from `X-Pagination-Next` header
4. Handle cursor expiration - cursors expire after **24 hours** (restart from beginning if error)
**Affected endpoints:** 38 action `/run/live` endpoints (e.g., `linkedin-extract-*`, `salesnavigator-*`, `recruiterlite-*`)
**Not affected:** Core API endpoints (`/runs`, `/schedules`) continue using `limit/offset` pagination.
### Improvements
* **Unified pagination**: Single cursor approach replaces mixed page/token system
* **Data consistency**: Cursors ensure stable pagination (valid for 24 hours)
* **Simpler implementation**: Always follow `X-Pagination-Next` header (no need to distinguish pagination types)
### API Parameter Updates
* `run_uid` path parameter now includes UUID format validation and description
### What This Means for You
* **Existing integrations continue to work** - `page` and `token` remain functional (backward compatible)
* **New integrations should use `cursor`** - recommended for all new implementations
* **Async/Schedule modes unchanged** - automatic pagination continues to work the same way
* **SDK users**: The `nextPage` helper still works - it returns the cursor URL
### 🚀 What's New
This release focuses on improved LinkedIn reliability, richer profile data, and new automation features.
### 🔒 LinkedIn Connection Reliability
* **Fixed**: LinkedIn accounts now reconnect smoothly without false authentication errors
* **Fixed**: Sales Navigator "upgrade required" errors no longer appear when visiting profiles within your daily limits
* Users should experience far fewer connection interruptions when running LinkedIn workflows
### 📬 LinkedIn Invitation Management
New actions for managing your LinkedIn invitations with precision:
| Action | Description |
| ------------------------ | --------------------------------------------------------------- |
| **Accept Invitation** | Accept a single invitation provided as input |
| **Withdraw Invitation** | Withdraw a single invitation provided as input |
| **Accept Invitations** | Accept all pending invitations (now supports async execution) |
| **Withdraw Invitations** | Withdraw all pending invitations (now supports async execution) |
These actions enable selective invitation workflows, giving you precise control over which invitations to accept or decline.
### 💰 Richer LinkedIn Profile Data
When searching LinkedIn Sales Navigator, you now get enhanced profile information:
* **Dynamic headlines** shown by LinkedIn
* **Job summaries** and descriptions
* **Work experience details** previously unavailable
This richer data enables better targeting and more personalized outreach.
### ⚡ One-Shot Callbacks for Automation Tools
New callback option for customers using automation platforms like n8n, Make, or Zapier:
* Receive **one notification when a job finishes** instead of multiple callbacks during execution
* Dramatically reduces costs for webhook-based integrations
* Set `callback.on` to `final` instead of `all` to enable this mode
### 📊 Better Run Visibility
* **Output counts** are now visible directly in run details
* Know exactly how many results each run produced
* Clearer ROI tracking and performance monitoring
### 🔐 Clearer Error Messages
When LinkedIn credentials are invalid, you now receive clear, actionable error messages instead of confusing codes. This makes troubleshooting faster and reduces frustration.
### 📝 What This Means for You
* **More Reliable**: LinkedIn authentication and Sales Navigator access issues are resolved
* **Better Targeting**: Richer profile data enables more personalized outreach
* **Lower Integration Costs**: One-shot callbacks can reduce automation platform costs significantly
* **Clearer Insights**: See exactly what each run produces with output counts
### 🚀 What's New
This release enhances Sales Navigator and LinkedIn actions with improved error handling, new callback fields, and standardized error responses.
### 🔔 Scheduled Run Tracking in Callbacks
Callback payloads now include a new field for scheduled runs:
```json theme={null}
{
"scheduled_run_uid": ""
}
```
This UUID field is included when the callback originates from a scheduled run, making it easier to correlate callbacks with their scheduled execution context.
### ✨ Enhanced Error Responses
We've significantly improved error response documentation across Sales Navigator and LinkedIn actions with detailed examples and new status codes:
| Status Code | Description | Change |
| ---------------------------- | -------------------------------------- | ------------------------------- |
| **400** Bad Request | Invalid parameters or input | Now includes detailed examples |
| **402** Payment Required | Access denied due to plan limitations | **New** |
| **422** Unprocessable Entity | Action aborted | **New** |
| **424** Failed Dependency | Integration or external service errors | Expanded with multiple examples |
| **429** Too Many Requests | Rate limit exceeded | **New** (replaces 500) |
#### Rate Limit Response Details
429 responses now include detailed rate limit information:
```json theme={null}
{
"error": {
"code": 429,
"status": "Too Many Requests",
"details": {
"Retry-After": "60",
"X-RateLimit-Limit": "2",
"X-RateLimit-Reset": "1754402101",
"X-RateLimit-Type": "/v1/actions"
},
"message": "Rate limit exceeded (/v1/actions)"
}
}
```
### 🔧 Bug Fixes & Improvements
#### LinkedIn Post Extraction
* 🐛 Sanitize LinkedIn post URLs by stripping query parameters to avoid duplicate/mismatched post references and improve cache/canonicalization correctness
#### Identity and Authentication
* ✨ If a `SN_ACCOUNT_UPGRADE` error occurs, the related identity will be ignored for 60 seconds, reducing repeated failures and improving overall action reliability (later, this will apply only to `managed` identities)
#### Cross-Action Error Messaging
* ✨ Standardized error messages across actions for consistency, aligning error codes/messages and improving client-side parsing
### 📝 What This Means for You
* **Better Debugging**: Detailed error examples help identify issues faster
* **Improved Rate Limit Handling**: 429 responses with headers enable smarter retry logic
* **Scheduled Run Correlation**: New `scheduled_run_uid` field links callbacks to their schedule
* **More Reliable Actions**: Automatic identity cooldown on SN upgrade errors prevents cascading failures
* **Cleaner Post URLs**: Post extraction now handles URLs with query parameters correctly
### 🚀 What's New
This release improves LinkedIn invitation handling, adds new output fields for post activity, and standardizes API error responses.
### 🔔 LinkedIn Invitations — Schema Improvements
We've refactored LinkedIn invitation actions for better consistency and reliability:
#### Standardized Invitation Identifiers
All invitation-related actions now use consistent identifiers:
* `linkedin_invitation_id` — String identifier for the invitation
* `linkedin_invitation_urn` — LinkedIn URN for stable reference
| Action | Changes |
| -------------------------------- | ---------------------------------------------------------------- |
| **Accept Invitations** | Simplified input — now requires only `linkedin_invitation_urn` |
| **Extract Received Invitations** | Renamed `sent_time` → `sent_date` for consistency |
| **Extract Sent Invitations** | Now returns `linkedin_invitation_id` + `linkedin_invitation_urn` |
| **Withdraw Invitations** | Output now returns invitation IDs instead of profile URL |
> ⚠️ **Migration Note**: If you're using invitation actions, update your integrations to use the new field names.
### ✨ LinkedIn Post Activity — Author Information
The **People Post Activity** action now includes detailed author information:
* `author` — The person who shared/posted the content
* `original_author` — The original content creator (for reshares)
Both objects include profile identifiers, details (firstname, lastname, headline), URLs, and images.
### 🔧 Bug Fixes & Improvements
#### LinkedIn Post Commenters
* 🐛 Fixed missing `first_name` and `last_name` fields — these were returning empty strings even when `full_name` was populated
#### LinkedIn Post Likers
* 🐛 Removed `linkedin_company_id` from output (field only applied to company likes)
#### LinkedIn Search Content
* ✨ Added `linkedin_company_url` to output
#### InMail & Message File Attachments
* 🐛 Fixed file validation for Base64-encoded attachments
* 🐛 Fixed filename extension validation regex
#### API Error Responses
* ✨ Added standardized error schemas: `PaymentRequired` (402), `Forbidden` (403), `NotFound` (404), `UnprocessableEntity` (422), `TooManyRequests` (429)
#### Pagination
* ✨ Updated default page size to 100 for improved throughput on applicable actions
### 📚 Documentation
* ✨ Added [`POST /runs/{run_uid}/cancel`](/v1/api/runs/cancel) — Cancel an async/schedule run
* ✨ Added [`GET /runs/{run_uid}/status`](/v1/api/runs/status) — Lightweight run status polling
* ✨ Added [`GET /runs/{run_uid}/outputs`](/v1/api/runs/outputs) — Poll for run outputs without callbacks
### 🚀 What's New
Major refactors landed around account rotation and integration limits, introducing caching layers that reduce load and standardize identity handling. HTTP service robustness improved (compressed responses, stricter content-length checks), and several LinkedIn actions received targeted fixes. Run callbacks now include additional identifiers/headers and OpenAPI updates, and error labels were standardized across the stack for clearer diagnostics.
### 🔔 Improved Webhook Delivery Identification & Deduplication
We're introducing two important enhancements to our webhook callback system to make delivery tracking and deduplication more reliable for all integrators.
#### 📌 New Stable Callback Reference (`callback_ref_uid`)
Webhook payloads now include a new field:
```json theme={null}
{
"callback_ref_uid": ""
}
```
This identifier is:
* **Stable across retries**
* **Unique per callback event**
* Designed specifically to allow **safe deduplication** on the client side
This ensures that even if your endpoint receives multiple retries, you can always identify which callbacks correspond to the same logical event.
#### 📌 New Delivery Attempt Headers
Webhook requests now include two additional HTTP headers:
| Header | Description |
| ------------------ | ------------------------------------------------------------------------------------------------- |
| **X-Callback-Ref** | Stable callback identifier, identical across retries (same as `callback_ref_uid` in the payload). |
| **X-Run-Callback** | Unique identifier for each delivery attempt. Changes on every retry. |
These headers allow you to:
* Track each individual delivery attempt
* Log retries more easily
* Distinguish between *event identity* (stable) and *delivery attempt identity* (per retry)
#### 📘 Why this matters
These changes bring our webhook behavior in line with industry best practices and make it easier to:
* Detect and safely ignore duplicate callbacks
* Correlate retries with the same logical event
* Debug delivery attempts via unique identifiers
* Keep payloads stable and idempotent
#### ✅ Summary
* **`callback_ref_uid`** is now included in the payload → *stable event identifier*
* **`X-Callback-Ref`** → stable ID in headers
* **`X-Run-Callback`** → unique per delivery attempt
* Retries always resend the **same payload**, preserving the original run status and data
📘 [Managing Callbacks Documentation](/v1/runs/callbacks)
### 🔧 Action Fixes
#### LinkedIn Job Search
* 🐛 Handled missing job data by throwing a descriptive action error, preventing silent failures.
#### LinkedIn Post Likers Extraction
* 🐛 Removed `linkedin_profile_handle` from the output schema because the source does not provide it.
#### Company Enrichment
* 🐛 Fixed company domain resolution during company enrichment to return the correct domain.
### 🚀 Edges New Release
This release focuses on improving platform stability, reliability, and user experience across LinkedIn automations, engagement identities, and core API functionality.
### 🔒 Improved LinkedIn Session Stability
* **Enhanced IP Management**: Implemented a new progressive IP rotation system that significantly improves LinkedIn session reliability
* **Fewer Connection Errors**: Reduced intermittent failures and reconnection issues during LinkedIn operations
* **Better Error Handling**: More reliable login, refresh, and checkpoint flows with improved retry logic
* **Smoother Workflows**: Users should experience fewer session-related interruptions and more stable automation runs
### 💰 Engagement Identities Billing Enhancements
* **Improved Billing Accuracy**: New buffer mechanism ensures more accurate usage tracking and billing for engagement identities
* **Multiple Plan Tiers**: Support for tiered pricing plans based on usage volume, providing better flexibility and cost optimization
* **Better Timeframe Validation**: Enhanced validation ensures billing stays within subscription timeframes for accurate charges
* **Clearer Error Messages**: Improved error messages when creating identities to distinguish between "No plan available" and "No more room in plan" scenarios
### 📊 Sales Navigator & LinkedIn Actions Improvements
* **Increased Limits**: Sales Navigator company employee search now supports up to **2,500 results** (previously 1,000) per request
* **Enhanced Profile Data**: Company logo URLs are now included in profile experiences when extracting people data
* **Fixed Post-Likers Actions**: Resolved issues with post-likers actions when using "share" URLs as input
* **Better Pagination**: Fixed error 424 issues on multiple-output actions by preventing unnecessary pagination when no results are available
* **Improved Post Commenters**: Enhanced validation for post commenters extraction to handle edge cases more reliably
### 🔧 API & Automation Improvements
* **Output Retrieval by Run UUID**: You can now retrieve outputs directly by Run UUID, making it easier to access run results
* **Better Workflow Support**: Enhanced workflow schemas with improved support for optional steps and output anchoring
* **Real Dates from IDs**: Actions now return accurate dates when retrieving data by ID
### 🛠️ Platform Reliability
* **Automated Data Retention**: Implemented automated cleanup of old runs and usage data to maintain optimal platform performance
* Runs older than 90 days are automatically archived
* Usage records older than 5 days are cleaned up
* This ensures faster queries and better overall system performance
### 📝 What This Means for You
* **More Reliable Automations**: LinkedIn actions should run more smoothly with fewer session interruptions
* **Better Billing Accuracy**: Engagement identities billing is now more precise and transparent
* **Increased Capacity**: Sales Navigator searches can now handle larger result sets
* **Improved Developer Experience**: Better error messages and API capabilities make integration easier
* **Faster Performance**: Automated data retention keeps the platform running smoothly over time
### 🚀 Edges New Release
New release adds several new features and improvements to the Edges platform.
* New [Engagement Identities](/v1/identities/engagement) mode for running LinkedIn outreach actions without consuming credits
* Improved UI overall with updated components for better fluidity and user experience
### 💰 Engagement Identities Pricing & Billing
* **New billing model**: Engagement Identities are billed monthly based on peak usage
* **Starting at \$7.99/identity/month** with dynamic pricing that decreases with volume
* **Credit-free outreach**: Run a comprehensive set of outreach actions (connect, message, follow, engage) without consuming credits
* **Flexible usage**: Convert existing identities to engagement mode or create new ones via API
* **Live monitoring**: Track current billing status and peak usage
📘 [Learn more about Engagement Identities](/v1/identities/engagement)
### 🚀 Update on Extract LinkedIn People
Here are the main changes:
* 👔 **With the experiences parameter**, we now retrieve only the first **five companies' experiences**. The total number may vary depending on the profile. If you want more experiences, use the [`linkedin-extract-people-experiences`](/v1/api/actions/linkedin-extract-people-experiences) endpoint.
* 💡 **With the skills parameter**, we now get only a **few skills (one or two)** — again, this can vary by profile. If you want more skills, use the [`linkedin-extract-people-skills`](/v1/api/actions/linkedin-extract-people-skills) endpoint.
* 📋 Both actions use the same input payload as `linkedin-extract-people`.
* ⚠️ **Note:** these two new actions share the same limit. We'll define the daily and per-minute request limits in a future release.
### 🔧 Technical Updates
* **Rate limit simplification**: Simplified rate limiting system (plans-based rate limits will be introduced in the next release cycle)
### ✨ New Action
* **[Extract LinkedIn Received Invitations](/v1/api/actions/linkedin-extract-received-invitations)** — New `linkedin-extract-received-invitations` action to extract invitations you've received on LinkedIn, including sender details, messages, and timestamps.
### 🔧 Improvements
* **Enhanced LinkedIn Native Login**: Improved stability with broader support for LinkedIn challenge handling, ensuring more reliable authentication flows.
* **Better Profile Data Extraction**: Enhanced "experiences" field parsing in the [`linkedin-extract-people`](/v1/api/actions/linkedin-extract-people) action for more accurate profile information.
* **Live Mode Enhancements**: `custom_data` is now included in both responses and errors for `live` execution mode, improving compatibility with automation tools and making it easier to track and reconcile requests.
### 🚀 Edges New Release
New release adds several new features and improvements to the Edges platform.
* New [Credit Consumption endpoint](/v1/api/workspace/consumption) to track credit usage
* New [Get Identity Limits endpoint](/v1/api/identities/limits) to track limits usage per identity and action
* New [Generate Login Links endpoint](/v1/api/identities/login-links) to create new login links for identities
* Improved documentation and examples in [FAQ](/v1/faq-troubleshooting#how-to-track-credits-consumption%3F), [Native Login](/v1/identities/add-manage-identities#native-linkedin-login-link), and [LinkedIn Limits](/v1/linkedin/limits#2-edges-smart-limits) sections.
### 🔔 Enhanced Integration Webhooks
* **New Event Type**: Added `AUTH_UPDATED` webhook event for monitoring LinkedIn account level changes
* **Enhanced Payload**: Webhook data now includes:
* `name`: LinkedIn account name
* `level`: Account type (Classic, Sales Navigator, Premium, Recruiter Lite)
* **Better Monitoring**: Real-time notifications when users upgrade or change their LinkedIn account levels
📘 [Integration Webhooks Documentation](/v1/identities/linkedin#webhooks)
### 📚 Documentation Improvements
* **API Consistency**: Updated endpoint descriptions from "List all" to "List" for better clarity
* **LinkedIn Limitations**: Added clear documentation about group member extraction limitations
* **Sales Navigator**: Enhanced filtering capabilities with updated OpenAPI specifications
### 🔧 Technical Updates
* Updated OpenAPI specifications with new Sales Navigator filters
* Improved API documentation consistency across all listing endpoints
* Enhanced error handling and limitation documentation
### 🚀 Edges v1 – Public Release
Edges is now publicly available — an API-first automation platform built to run precise, scalable and fully observable actions on top of platforms like LinkedIn and Sales Navigator.
***
### 📦 TypeScript SDK
The [Edges TypeScript SDK](https://www.npmjs.com/package/@edgesrun/sdk) offers a type-safe and minimal way to interact with the Edges API.\
It handles authentication and exposes all endpoints for:
* Action execution (`live`, `async`, `schedule`)
* Identity & integration management
* Run tracking & callbacks
***
### ⚙️ New API Core
Edges exposes a fully documented and production-grade REST API with:
* Modular **actions**, identified by scope (e.g. `linkedin-extract-people`)
* Modern **execution modes**: `live`, `async`, `schedule`
* Strong separation between **members** (dashboard users) and **identities** (automated accounts)
* Robust **pagination**, **rate limits** and **error tracing**
📚 [Explore the full API Reference](/v1/api/)
***
### 👥 Identity-Centric Design
* Identities represent your automation accounts (e.g. a LinkedIn profile)
* Create and manage identities independently from workspace members
* Use native login or cookie-based authentication
* Monitor integration status with webhooks
📘 [How to connect a LinkedIn Identity](/v1/identities/linkedin)
***
### 🔁 Async, Schedule & Callbacks
* Run **bulk actions** at scale via the `async` mode
* Program runs to start **later** (`schedule_at`) or **recurrently** with `cron` via the `schedule` mode
* Provide a `callback_url` to get results pushed when ready
* Enrich callback payloads with your own `custom_data` to simplify reconciliation
📘 [Execution Modes Overview](/v1/runs/overview) & [Async Callback Guide](/v1/runs/callbacks)
***
### ✉️ Featured Actions
* **LinkedIn** Authentication, Extract, Search, Find, Intents, Messaging & Engage
* **Sales Navigator** Search, Metrics & Engage
* **Recruiter Lite** Search
***
### 🧠 Developer Experience First
* Full OpenAPI-based, comprehensive documentation
* Instant authentication with API keys
* Built-in account rotation logic and identity modes
* Tracking and observability with run UIDs, callbacks and callbacks api
* Fast iteration cycles
📘 [Getting Started Guide](/v1/quickstart)
#### LinkedIn Post Activity
* **Added** [Understanding Author Fields](/v1/api/actions/linkedin-extract-people-post-activity#understanding-author-fields) for **Extract LinkedIn People Post Activity**:
* `activity_author` — Person who performs the activity (e.g., reshared).
* `author` — Author of the first-level (visible) post.
* `source_post` — Source post (2nd level) when the visible post is a reshare; includes `author`, `linkedin_post_id`, `linkedin_post_url`, `content_text`; `null` when there is no 2nd level.
* Examples for original post, reshare without comment, reshare with comment, and reshare of a post that shares another post.