Edges Overview

Edges is a REST API for LinkedIn and Sales Navigator data and actions, sold to developers building GTM software and AI agents. One endpoint pattern runs every action live, in the background or on a schedule, and an action runs either on accounts Edges operates or on your users’ own LinkedIn accounts, which Edges calls identities. It bills on two meters: credits for each record returned, and a monthly fee for each identity that sends invitations or messages. Edges enforces its own per-identity daily limits and ramps a new account up over five days.

What Edges is

Edges calls itself “The LinkedIn API, built for builders”, and its machine-readable summary puts it as “one REST API for LinkedIn and Sales Navigator data, search, messaging, engagement, and timing signals”. The audience on the homepage is people who “power their software” with it: GTM products, data pipelines and agents. It covers two kinds of work through one key. Reads return people, company, post, job and search records. Writes send connection requests, messages and InMails, visit profiles, like and comment, and accept or withdraw invitations, from an account that is connected.

It is a developer product with no campaign app on top. Your code decides what runs and when, and Edges executes it on its own infrastructure. The company behind it is Edges Labs LLC, a Delaware limited liability company, under its terms of service. Every page carries the line “Edges is not related to LinkedIn and is not an official LinkedIn product.”

How an identity connects and runs

In Edges’ model an identity is “an end user in your app”, and a LinkedIn account connected to it is an integration. Actions run in one of two ways, and the choice decides cost, capacity and which actions are possible.

• Your users’ own accounts. Connected through Edges’ Chrome extension, a native login with email and password where Edges handles LinkedIn’s checkpoints, or session cookies. Edges can also issue a login link for each identity, valid for 48 hours, so the user connects without your app collecting the password. Messages, invitations, InMail and profile views need this mode, because they “must be attributable to a specific person”.
• Managed accounts. Set identity_mode: "managed" and the action runs on accounts Edges provisions and operates, “through third-party accredited services”, with no connection step and Sales Navigator search on Edges’ seats. It costs 1.5 times the standard credits, “plus a possible monthly cost at higher volumes”, and it cannot send anything or read data that belongs to one account.

A third setting, auto, spreads work across every connected identity in the workspace. Each identity gets its own IP address, provisioned in the country of the identity’s time zone, and Edges asks for that zone to match where the person really logs in. Supported licences are LinkedIn Basic, Premium Business and Sales Navigator Core, with Premium Career listed as not compatible. Session events such as AUTH_EXPIRED arrive by webhook, and runs blocked by an invalid account can be recovered after it reconnects.

Edges pricing

Edges bills two meters on one invoice. “Credits meter your data”, one credit per record returned, and “Identities meter your outreach”, one per LinkedIn seat that sends invitations or messages. “A pure-data customer pays $0 on identities. A pure-outreach customer pays $0 on credits.” Each meter has one rate for the whole volume, set by the band the volume falls in. The figures below are the calculator’s, in US dollars.

The two meters meet in one place. On an identity, “Sending messages and connection requests or extracting messages and conversations are free”, while enrichment still costs credits. Outside that mode, sending a message costs 1 credit. A failed run costs nothing: “you’re billed on success”. Credits reset every month and overage is charged at the current tier’s rate. The free tier is “100 free credits and 1 identity to try every endpoint”, with no card, renewable each month for personal use. Annual terms take 8% off a one-year commitment, 12% off a one-year prepayment, 10% off a two-year contract and 15% off a three-year one.

The pages disagree on the starting prices. The calculator in dollars starts credits at $43.99 per 1,000 and identities at $8.79. The page’s own description, its schema data and the Engagement Identities documentation give “$39.99”, “$1.49”, “$7.99” and “$3.82” with a dollar sign, which are the calculator’s figures in euros. The euro switch on the calculator shows them as €39.99 and €7.99.

The limits Edges enforces

Edges caps LinkedIn activity itself. “Smart Limits are the allowances Edges enforces on each LinkedIn action, per identity”, and its documentation describes them as set “below what LinkedIn tolerates”. Each cap is a rolling window per identity and per action, so reaching the invitation cap leaves the identity’s other actions running. The 24-hour quotas for the main actions on an account without a paid subscription:

Most actions also carry a per-minute cap, and some add 15-minute, hourly and four-hour caps. A newly connected identity starts lower: “Full capacity is reached after 5 days”, at 20%, 40%, 60%, 80% and then 100% of each 24-hour quota, so day one allows 5 connection requests. “Ramp-up is on by default for every identity”, and a custom limit on the integration skips it. On Scale and Enterprise a customer can request an override of the limits.

What happens at a cap depends on the mode. “In live mode the request fails with a limit error. In async and schedule modes the work is postponed and resumes automatically once the window frees up.” The endpoint GET /v1/identities/{identity_uid}/actions/{action_slug}/limits returns the number that applies to one identity right now, ramp-up included. Edges adds that it “counts only the actions performed through the platform”, so manual activity and other tools still count against the account with LinkedIn. LinkedIn’s own published numbers are compiled on LinkedIn automation statistics.

The Edges API

The API lives at https://api.edges.run/v1 with the key in an X-API-Key header. “Canonical call: POST /v1/actions/{slug}/run/{live|async|schedule}“, so one action runs synchronously, as a background job with a callback, or on a cron schedule.

• Size. Edges says “70+ actions”. Its OpenAPI catalogue lists 86 action slugs, counted on 13 November 2026, four of which are outside LinkedIn: three Google Maps actions and an email enrichment action.
• LinkedIn sends. Connection requests, messages, InMails, event invitations, likes, comments, follows, profile visits and invitation withdrawals, in standard LinkedIn and Sales Navigator versions where both exist.
• Rate limits. Separate from the per-identity caps, the plan sets the request rate: 1 live request a second on the pricing page’s Pilot column, rising to 20 to 40 on Enterprise. The rate-limit documentation names a Trial tier at 1 a second and 60 a minute and Bronze at 2 and 120.
• Clients. An official TypeScript SDK, @edgesrun/sdk. Python goes through plain HTTP.

Callbacks deliver async and scheduled results, with a replay endpoint for failed deliveries, and the identity endpoints create, list and delete identities and connect or disconnect their LinkedIn accounts.

The Edges MCP server and Agent Skills

Edges runs a documentation MCP server at https://docs.edges.run/mcp, with three tools: search, browse and send feedback. It is explicit about its scope: “This is a documentation server. It reads these pages and nothing else – it does not call the Edges API, does not run actions, and never sees your API key.” No key or account is needed.

To have an agent call the API itself, Edges ships Agent Skills instead: “76 curated Agent Skills” under the Apache-2.0 licence, 68 of them for actions, installed with npx skills add edgesrun/linkedin-skills or as a Claude Code plugin. The skills page describes them as “Cookieless by default – no LinkedIn account required”, which points them at managed reads. The agent then calls the REST API with your key, so any send runs through the same identities and limits as a call from code.

The shape of job it fits

Edges fits a developer whose product reads a lot of LinkedIn data and also needs some outreach from its users’ accounts. The data meter bills per record and nothing for an empty result, the managed mode reads without any connected account, and the identity meter makes outreach a flat monthly cost per seat with sends free on it. The Smart Limits table, the five-day ramp and the per-identity limits endpoint are documented in detail, and the background and scheduled modes wait out a cap instead of failing.

The limits of the design follow from the same choices. The minimum data plan is 2,500 credits a month, about $110. The managed pool’s monthly cost at volume is set with sales and not published. An agent gets the API through Agent Skills and a documentation server, with no MCP tool that acts on LinkedIn. And the starting prices read differently in euros and in dollars across Edges’ own pages.

Edges alternatives

GTM API is our own product, so it is listed first. The others are described from their own pages, in alphabetical order. They split by what is metered: records of data, or connected accounts that act.

Read the table by what you pay for. For LinkedIn data at volume with no account of your own, Edges’ managed mode and Captain Data bill by the record. For sending from your users’ accounts, Edges bills per identity, and Linked API, LinkupAPI, Unipile and GTM API bill per connected account, with LinkupAPI and Unipile leaving daily limits to the caller and Edges, Linked API and GTM API applying them on the server. How LinkedIn APIs price their calls is laid out on LinkedIn API pricing.

Frequently Asked Questions

Sources & Further Reading

• Edges, homepage · pricing (calculator in US dollars, identities at ?mode=outreach) · llms.txt · Agent Skills · OpenAPI catalogue · terms of service, read 2026-11-13
• Edges developer documentation, core concepts · identity modes · connecting a LinkedIn identity · engagement identities · Smart Limits · identity ramp-up · plan and API rate limits · documentation MCP, read 2026-11-13
• Related: LinkedIn API pricing · Best LinkedIn API · LinkedIn automation API · Unipile overview · LinkedIn automation statistics

Michael SalamatovGTM engineerRuns the gtm-api.com blog. Covers LinkedIn limits, account health and the API side of outbound.in

Last updated: November 2026 · Definitions and benchmarks verified at time of publication