Changelog
Hand-curated notes for every release of the Pricing Compute Network. New capabilities, API changes, security hardening, design work. If it is here, we thought it was worth your attention.
Last 4 weeks
Releases each week, by category
| Week of | New | Improved | Fixed | Faster | Security | API | Design | Docs | Total |
|---|---|---|---|---|---|---|---|---|---|
| Sep 28 | 25 | 22 | 28 | 0 | 10 | 18 | 12 | 9 | 124 |
| Sep 21 | 9 | 9 | 9 | 0 | 3 | 1 | 0 | 1 | 32 |
| Sep 14 | 6 | 0 | 7 | 0 | 1 | 0 | 0 | 1 | 15 |
| Sep 7 | 7 | 4 | 5 | 0 | 1 | 5 | 1 | 1 | 24 |
| Jun 1 | 7 | 0 | 3 | 0 | 1 | 0 | 0 | 0 | 11 |
| May 25 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 2 |
| May 11 | 9 | 1 | 0 | 1 | 2 | 1 | 2 | 1 | 17 |
| Apr 13 | 1 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 2 |
Every release, by category
Releases for people, for agents calling the API, or both
25 days with a release. Busiest: Oct 2, 2026 (44)
Release days: 25 days with at least one release. The table under the weekly chart lists every week.
Hover a day, or step through release days
When you report whether a quoted price was accepted or rejected, Last Price now learns how sensitive your buyers are to price, per product and for the whole workspace, and the default optimizer can price from it instead of a generic guess.
An LLM prompt or ML model pricing function can now run on an inference provider your workspace connected, with your own model and pricing instructions. Before, these functions quietly used the basic rule-based formula.
The Notifications panel in the sidebar now lists approval requests, failed webhook deliveries, email to your agents, disputes and proposed price changes for the selected workspace, and the Monitor tab shows how many are unread.
A Value Receipt now shows what an agent cost, what its work would have cost done by a person, and the ROI, all measured from your own data over the last 30 days. Agents and your code can read the same numbers from the API.
The menu, footer and page heading now say Changelog, matching the address, so the name you read is the name you type.
Owners and admins now get an email when a call is held for approval, a dispute opens, or a webhook endpoint keeps failing, and choose which ones in their profile. Agents get the same alerts through the notifications feed, MCP, webhooks and their own inbox.
The featured notes at the top of the release page are now chosen by hand, so the headline changes are the first thing you see.
You can now create your API key, check for your first call and activate your pricing function without leaving Get started, and any page a step does open leads straight back.
Learned pricing no longer needs you to report every quote. Quotes nobody reported on now count too, read from the orders that followed them, so an estimate is ready with far fewer reports and is less skewed by which outcomes buyers bother to report.
Every code block and command snippet across the docs, developer pages, and dashboard now has a Copy button so you can grab any snippet without selecting text by hand.
A new pricing analysis reads your workspace's own records and says, in plain sentences with the numbers behind them, what is ready, what is stuck and what to do next. Free, for people and agents alike.
Charm endings no longer push cheap prices up, sub-cent prices never round to zero, and a price is never below the unit cost you send unless you save guardrails that allow it.
Every release note now links directly to the relevant documentation section and to the screen in the app where the feature lives, so you can go from reading about a change to using it in one click.
Screenshots on the release page now fill their frames completely, with no grey borders around them.
A new Releases page lists every deploy to production, newest first, with the date and the changelog notes it shipped.
Three new webhook events tell your own tools when a person has something to act on: an agent call waiting for approval, a dispute opened, and a price change proposed. Route them to a chat channel, a pager or a ticket queue.
Competitor price observations you load for a pricing cycle, and checkout sessions you retry with an Idempotency-Key, are now kept strictly to your own workspace.
The agent inbox now falls back to locally stored messages when the external mail service cannot be reached, so mail remains visible even during an outage.
The pricing step of agent setup now asks what each fee costs: an amount, a currency and how often it is billed. It used to record only that a fee was switched on, and the agent's page now shows the fees you set.
Banned accounts now appear as on hold in staff pages, with no approve or reject actions that could trigger emails. Repeat approvals no longer resend the welcome email.
Customers, payments, invoices, orders and the other billing pages keep their messages readable on phones, Orders can export to CSV, and the event tracking and cost tracing guides show your real setup instead of a demo.
Fetch company info in Company settings now fills in your company's name, description and logo from your homepage. Before, it only said that company enrichment was not configured.
Two pricing functions that were listed but could not be called now work: a significance test that says whether a price variant has won, and a traffic allocator that decides how to split the next visitors.
A cost multiplier you set on a function is now kept, and built-in functions bill at their listed multipliers, from 1x to 5x. Until now every call was metered at 1x, whatever the function said.
If Google sign-in or sign-up cannot start, the page now says so in plain words and offers email instead of showing technical setup details.
Connecting an inference provider on the Compute page now saves the key for your workspace, encrypted, after the provider confirms it works. Before, the card said Connected but the key was thrown away.
When you report whether a quoted price was accepted or rejected, Last Price now learns how sensitive your buyers are to price, per product and for the whole workspace, and the default optimizer can price from it instead of a generic guess.
Seven new MCP tools let an agent read and change one of its pricing functions, turn outcome pricing and balance billing on or off, set the auto top-up offer, and read its outcomes, usage and priced calls, without the dashboard. The LangChain, AI SDK and Agents SDK tool packages offer them too.
An LLM prompt or ML model pricing function can now run on an inference provider your workspace connected, with your own model and pricing instructions. Before, these functions quietly used the basic rule-based formula.
The Model Profiling page no longer shows six sample backends, a sample GPU and a sample latency histogram. It reads your profiling server, and says so plainly when none is connected.
Model Profiling now runs on the AITune server your workspace connects on the Compute page. Before, it could only use a server set up by Last Price, which did not exist, so the page always said AITune was not connected.
The Notifications panel in the sidebar now lists approval requests, failed webhook deliveries, email to your agents, disputes and proposed price changes for the selected workspace, and the Monitor tab shows how many are unread.
Confirmations and error messages now appear on screen after you save, delete or send something in the dashboard. Before, they were raised but never shown, so a save or a failure gave no feedback.
Every release note is now tagged with the parts of the product it changed, such as Webhooks or Credits and billing. Pick a feature to see only its notes, or tap a tag on any note.
Ask for a receipt for any priced call by its id, or for a statement of many by date range, as JSON, CSV or a printable page. Receipts are made only when you ask, from the dashboard, the API or two new MCP tools, and a statement counts the value signals you set up for each agent.
Settings, Sales playbook now keeps your ideal customer profile, value propositions, pricing guardrails and commercial terms for the workspace. They used to disappear when you left the page.
Rate limits on dashboard sessions now track individual people rather than shared network addresses or workspaces, so teammates on the same network no longer affect each other's limits.
Receipt statements now show each agent's human value and ROI alongside its costs, using the same figures as the agent's Value Receipt.
The built-in Usage table, Invoice table, Payments table, Container and Pricing calculator blocks were placeholders. They now render real content: the tables read your own data through a small server route you copy into your app and protect with your own sign-in, and the calculator works out graduated tier pricing.
Settings, Tax now keeps your company address, US nexus states, EU VAT (OSS) number and other VAT or GST registrations. They used to disappear when you left the page.
The header names the page you are on, the sidebar shows every section and highlights where you are, and pages that showed sample or placeholder content now show your own data or say plainly what is not available yet.
The profile photo and company logo uploaders now say which image types and sizes they take, and no longer name the service behind them.
The usage signals endpoint now accepts filters for customer and event type, letting you retrieve exactly the signals you need without post-processing the full list.
A Value Receipt now shows what an agent cost, what its work would have cost done by a person, and the ROI, all measured from your own data over the last 30 days. Agents and your code can read the same numbers from the API.
Settings, Value receipts no longer offers five sample agents with sample signal counts, and a configuration you create is now saved. It lists your workspace's agents and the signals each one has actually recorded.
Every webhook now carries its signature in x-lastprice-signature and its timestamp in x-lastprice-timestamp. The previous signature and timestamp headers are still sent with the same values.
Status badges, code examples, and small labels are easier to read in both light and dark mode, the show-password button works with screen readers, and the dashboard now respects your system's reduce-motion preference.
The API reference now fully documents each agent's kind and inbox visibility, giving callers a complete schema to build against.
A new docs page walks a person through setting up a dedicated wallet so their AI agent can pay for each price computation in USDC with x402, with no account.
The API reference now documents outbound webhooks, webhook deliveries and test pings, reading a single routing policy, updating an agent, recording costs, pricing models, listing usage signals, the tester tag, account deletion, and the two machine-readable discovery documents.
API reference screenshots embedded in release notes now render in dark mode, matching the surrounding page theme.
A new Approvals screen lists the agent calls waiting for a person's sign off, with Approve and Reject buttons, and the approval link an API call returns now opens it.
Tell Last Price who pays for model inference. When your customer brings their own AI subscription or key, the model cost drops out of the price, so you can offer a cheaper plan and keep your margin.
All release notes on the changelog page now appear as a grid of cards, making it easier to scan what has shipped at a glance.
Highlighted release notes on the changelog page now sit side by side in a row and scroll horizontally on smaller screens.
The Providers tab on the Compute page no longer shows sample providers with made up token counts, uptime and latency. It lists the pricing functions your workspace can route to, with this month's real requests and cost.
Compute no longer ignores a function named by name or fails with a server error on a catalog entry that cannot price, session-only routes say why an API key is refused, and the API reference documents the payment headers and Rosetta's sandbox limit.
The default pricing function now returns a profit-maximising price that moves with your cost and demand data, instead of always recommending the bottom of the allowed range, and the elasticity function no longer raises prices when a test shows customers are price sensitive.
The new credit bundle form no longer reports a bundle as created when nothing was saved, and Batch Compute shows an error instead of sample results when a request fails.
The tabs on a customer's page now show one section at a time, and Payments, Credit balances and Costs list that customer's records.
Code samples in developer documentation are now readable in light mode, and small links across the site meet a larger tap-target size.
The documentation pages are readable without signing in, and AI agents are directed to llms.txt for the current access rules.
The Experiments page no longer starts from sample tests with made up numbers. It lists your workspace's experiments and their real results, and says clearly when it is loading, empty or could not load.
The changelog entry describing how function credentials are encrypted at rest was corrected for accuracy.
Authentication configuration stored with each function is now encrypted at rest, so credentials are protected even at the storage layer.
The Get started page now shows three real first steps (create an API key, compute your first price, publish a pricing function) and ticks each one off once your workspace has done it.
The JSON and RSS release-note feeds now carry the 50 newest notes, so agent fetch tools can read them, and signing in to list a function says why an account is needed.
Two new MCP tools, get_price_guardrails and set_price_guardrails, let an agent read and replace its workspace's price guardrails. The LangChain, AI SDK and Agents SDK tool packages offer them too.
Usage analytics for MCP tool calls now include the failure reason and the name the client reported, making it easier to diagnose integration problems.
Settings, Billing now has an Outcome pricing section: turn it on, see how many priced calls were used or not, what came back to your balance, and mark a recent call as used or not used.
Turn on outcome pricing and every compute call paid from your prepaid balance opens a 7-day window. Report the price as rejected in that window and 70% of it goes back to your balance.
Save a maximum change, a minimum margin, a never below cost rule and the optimizer's search range once, and every price your workspace is given respects them, whichever pricing function computes it.
The privacy policy now explains that last-price.ai keeps daily counts of API and MCP access. These counts never include full API keys, tokens, or IP addresses, and are deleted after 180 days.
The privacy policy now explains the product analytics and session recordings we may use, and that typed text, keys and secrets are masked.
Release notes can now carry a screenshot of the change, shown on the card and in full on the note's own page.
Web and email addresses written in release notes are now automatically turned into clickable links.
Release notes now carry a screenshot of the change they describe, including an API reference screenshot for API-related notes.
Opening an expired or already used password reset link now shows a clear message and a button to send a new link, instead of a form that fails with a session error.
A saved default routing policy could be skipped for some pricing requests; it now applies to every request, wherever it is served.
The Price Intelligence page no longer shows sample scrapers, sample pool prices or a made up credit balance. It shows the prices your workspace has contributed, your real credit balance and history, and queries the shared pool only when you ask.
Search engines can now discover every public product page, including Rosetta, PriceRouter, UCP, agent wallets, and Price Check.
The staff activity page now shows the last seven days of API and MCP access, broken down by client, tool, and key.
Visiting the staff area no longer shows a page-not-found error. It now lands directly on the activity page.
The staff release page now lists every merged change that has not yet reached production, with no extra setup required.
Staff user detail pages were showing an error instead of loading; they now display correctly.
Navigation, sign-in, and call-to-action controls across all public pages are now at least 44px tall, making them easier to tap on phones.
Usage outcome reports larger than 4 KiB now receive a 413 response, and the public function catalog returns a retryable 503 when its rate limit cannot be checked.
People who sign up while their account still needs approval now get a "You're on the list" email instead of one saying their account is live.
The summary on the agent wallet guide now shows a single address, the wallet that receives payments, and names the currency as USDC with a link to verify the official token.
The release notes page now charts everything we ship: a weekly chart by category, the overall mix of work, who each release is for, and a calendar of release days.
The marketplace price history source now reads the regional store you configure and reports its prices in that store's currency, so catalogs priced in euros, pounds, yen and other currencies get competitor matches.
The credential a function sends to its endpoint is now write-only. No API response, key or dashboard view returns it; responses say whether one is saved.
The credential lookup now reports the actor kind behind every credential, so an agent's own API key answers agent instead of null.
The public function catalog and the public marketplace listings can now be read with no API key, so an agent can see what the network offers before it signs up.
The public pipeline run operation, which has never computed a price and always answers 501, is now marked deprecated and says so in its response headers.
A key or agent token limited to specific scopes is now refused on routes that take no scope, and pricing function, routing policy, batch and UCP routes now name the scope they need.
The support address now opens a real page with the support email, links to the docs and agent discovery files, and fixes for the most common errors.
Nobody, a co-owner included, can demote or remove the workspace owner from the team, and no change can leave a workspace without an owner. Suspended admins can no longer change the team.
Recording a usage signal takes eventType, agentId, customerId, properties and metadata at the top level, and the reference and dashboard docs now say so.
In the dashboard, viewers and billing members can no longer create API keys, and only owners and admins can manage webhooks, store connections and workspace settings, while only owners, admins and billing members can move money.
An access token an agent gets by exchanging its key now reaches exactly what that key reaches, so it can read only its own agent and cannot call the routes meant for people.
A workspace with balance billing on now pays for batch pricing from its prepaid balance, one call per item at the prepaid price, and a batch the balance cannot cover in full is refused with nothing charged.
You can now change the events an existing webhook endpoint receives from the dashboard, including the product updates event, without recreating the endpoint or its signing secret.
Agents can send a competitor offer and business limits to get a metered price suggestion or a clear request for missing inputs.
Unsubscribing from one of our marketing emails now turns off the Marketing emails switch in your profile, and every change to that choice is kept in a history you can see.
A tool call on the MCP endpoint that runs past 30 seconds is now stopped and answered with a JSON-RPC timeout error that says to retry, instead of the connection ending with no answer.
What shipped is now also published as a newsletter every other week, in an edition for people using Last Price and a more technical one for agent builders, with a public archive and an RSS feed.
Compute calls paid per call, taken from a day pass or drawn from a prepaid balance no longer count toward a workspace's billable usage cost, and the usage page shows the split.
A self-registered agent workspace can buy a day pass with one wallet payment, adding extra compute calls for 24 hours that are used once its daily limit is reached.
Owners and admins can give an agent a key with the Manage the team permission, so it can invite people, change roles and remove members for them, within strict limits and with every change listed on the Team page.
A self-registered agent that reaches its daily compute limit can pay for the next call with a wallet payment and have it run straight away in its own workspace, with no claim and no person.
The API reference now includes the scripts and styles it needs to display the current API specification.
The agent manifest, the MCP manifest, llms.txt and the self-registration response now say where the changelog feeds are, how to keep only the entries written for agents, and when to check them.
Calls made with an API key to the pricing, function list and credential routes now carry up to three notices about changelog entries for agents the key has not been shown yet, and the API keys page shows when each key was last used.
New store connections can import competitor offers for pricing proposals while preserving price limits and approval rules.
Every successful price computation now reports what it saved against the most expensive eligible function, up to three next steps, what is left of your limits, and a receipt for the call.
The Legal page and the Terms of Use now explain how to report content on Last Price that you believe infringes your copyright.
A retiring API operation now says so on every response with standard Deprecation, Sunset and Link headers, and a new platform.changelog webhook event tells your endpoint when a changelog entry for agents is published.
The documentation guide loads available access options and limits from the current deployment and explains key rotation and request tracing.
The errors an agent meets most now carry a hint, the next call to make, a valid example body, or the wait before retrying, beside the fields they always had.
A Marketing emails switch in profile settings takes you off our marketing list, and our emails can now carry the sender's postal address.
The hosted MCP endpoint now implements the standard Streamable HTTP transport, so a client can hold a session and open a stream that tells it what changed, while plain JSON-RPC POST keeps working exactly as before.
A workspace can top up a prepaid balance in USDC with one wallet payment and have compute calls draw it down at a lower price than paying per call, with a bonus on larger top-ups.
Pricing conversations ask clear questions, retain your answers and explain assumptions before you run a computation.
The Terms of Use now say you must be at least 18 to use Last Price, and the sign-up and sign-in pages link both the Terms and the Privacy Policy.
Keys can now be limited to billing or commerce work, those routes check the scope, and every commerce operation is in the API reference.
The experiments page and function catalog now describe the two-arm Experiments API, discovery says which credential reaches workspace and team admin, and wallet sign-in errors match every other sign-in error.
GET /api/usage/llm/events lists individual AI-backed calls, and PATCH /api/usage/llm/events/{id}/outcome records whether one was accepted, alongside a quality score.
The AI tab on the Usage page now charts tokens per day across metered AI operations and shows what they cost in euro next to the dollar figure. AgentMail calls are not included yet.
GET /api/public/leaderboard's confidence component now averages the confidence behind each listed pricing function's actual requests, instead of staying permanently unmeasured.
A self-registered sandbox key can rotate itself, discovery always states the limits in effect, a claimed workspace becomes an ordinary one, and a pay-per-call payment whose settlement was refused can be presented again.
The dashboard address now opens the app, the commerce endpoint agents discover answers, and demo errors report the remaining demo budget.
A workspace an agent creates for itself is now on its own Agent Starter plan, with 25 compute calls a day until a person claims it.
A claimed agent workspace now sits under the workspace you pick, keeps everything in it, and a trusted agent can claim for you with a new permission.
When you claim a workspace an agent created, the key the agent holds is now revoked unless you choose to keep it, and the claim page says plainly what keeping it means.
The public leaderboard counts each consumer once however many workspaces they use, and no longer counts a provider's other workspaces as consumers of its own function.
GET /api/public/leaderboard's adoption component now counts the distinct workspaces that actually use each listed pricing function, not a placeholder.
People can sign in or sign up with a Base wallet, link wallets to an existing account, and see the pay-per-call payments made from them.
Price verification accepts price as an alias, the API description is available as JSON, and the contract now describes what the product does.
The no-key pricing demo is now listed in llms.txt, agents.json and the docs, the homepage example shows the demo's real answer, and every advertised discovery link resolves.
An agent can now register itself and receive a sandbox workspace and a restricted API key immediately, plus a claim link for a person.
Key management under /api/api-keys now needs a person's session; a scoped token cannot mint a key wider than itself.
Every 401 response now carries a code, message and retryable flag, so agents can handle auth failures the same way everywhere.
The public marketplace and the Elo Index now list eight of Last Price's own built-in pricing functions, marked as built-in.
Live traffic now names the agent or client behind every pricing request and separates your own agents, your testers and everyone else.
The Browse tab in the dashboard marketplace lists the functions actually published on the network instead of sample entries.
POST /api/mcp/rpc serves the Last Price MCP tools over HTTP, so agents that configure themselves from /.well-known/mcp.json can connect.
OAuth client credentials frameworks can now find the authorization server metadata and register a client with no human.
Sessions and recent account activity now show only real data, not sample devices.
DELETE /api/user/sessions answers 400 for a currentSessionId that is not a session id, instead of a 500.
Opening settings with an expired session now takes you to sign in and back, instead of showing an empty page.
Clicking Invite user and typing in the invite form no longer hold up the page.
Calls to /api/tenants, the experiment and embed routes, and Rosetta ingest and translate now return a JSON 401 when credentials are missing, instead of an HTML 500.
Where enabled, a call to POST /api/compute/price with no credential answers 402 with x402 payment requirements, and a signed USDC payment buys that one call.
Direct execution of the internal authentication verification helper is now restricted to authorized callers only, reducing the surface area for privilege misuse.
Every operation now states whether it is open, needs an API key or JWT, or works only from a signed-in dashboard, and the translation API is documented.
The staff activity page and signup alert emails now report confirmed and pending signups separately instead of a single total.
An external pricing function that answers with no usable price or confidence now counts as a failed attempt and falls through to the fallback chain, instead of reporting your own price back as a computed answer.
Every computed price now explains how it was produced, and a new live traffic view shows your requests as they happen.
You can now record whether a computed price was accepted or rejected, plus an optional quality score, against the usage event that produced it.
Staff can inspect account activity and email alert status, while agents get a direct signup guide.
Agent inboxes can send new messages and reply within the original email thread.
Malformed experiment responses now report unavailable instead of generating prices from fallback assumptions.
Items shared into the mobile app at launch stay available while your saved login is restored.
Verified accounts can start without staff approval, with an additional agent identity sign-in option on configured deployments.
Multi-modal pipeline runs no longer return simulated prices as if they were computed; authorized execution reports 501 until real stages are wired, while pipeline discovery keeps working.
General workspace settings no longer return or replace stored integration credentials.
Routing to a specific function checked that it existed and was active, but never that the caller was allowed to reach it, so a workspace could compute against another workspace's private function by naming its id. Every compute path now allows only built-ins, your own functions, and functions listed publicly on the marketplace.
The Last Price MCP server now responds to tool calls before a credential is configured, so agents can explore capabilities without a sign-up step.
Every place the origin allowlist appears now explains what it does, and the same-origin documentation is attached to the correct function.
Agents can now request a price for work that has never been priced before, and no longer need to supply a product ID or category to get an answer.
A new verify endpoint lets you (or your agent) ask whether a quoted price is reasonable for its context, and it now uses a real reference price with a clear disclosure when evidence is not independent.
Single-price requests enforce the compute grant, and successful router calls are recorded in usage history.
Customers can now carry contacts, the people an invoice is addressed to, with a full set of endpoints and a place to manage them in the dashboard. An order can name one of them as its billing contact, and the getting-started walkthrough now shows the calls this API actually accepts.
A routing policy's strategy field was accepted, validated, and stored, then never read, so a policy set to cheapest sent ordinary traffic to the head of its fallback chain instead. The router now honours it, and a policy that could never be carried out is refused when you write it. Read the note on cost and latency ceilings: existing policies can start routing differently.
Usage now has an AI tab showing calls, tokens, and cost for agent generation and task estimation, including which key paid for each call.
The dashboard now shows a dedicated AI usage tab with date-filtered breakdowns, and usage data errors no longer surface raw database messages.
Filtering functions by built-in status works, variant metadata can shape the pricing response, and the marketplace commission default is 30% in every place it is stated.
The interactive API reference no longer reports a missing specification, and a misconfigured one now reports the real error.
Nine endpoints in the published reference described calls that the service does not answer. The reference now names the right method and request body, and the endpoints that were never built have been withdrawn.
Customer, order, invoice, payment, dispute, credit and cost responses are documented with the fields and types the API actually sends, so a generated client matches what comes back.
Ask Last Price switched from sample output to live AI-generated responses, and each message now produces a downloadable artifact.
Removed a token generator that called an endpoint which no longer exists, and corrected the authentication guidance it was part of.
Draft and cancelled orders can now be deleted from the orders list or through the API, and their line items go with them. An order that has started billing cannot be deleted outright: cancel it first.
The Elo Index is a public leaderboard that ranks pricing functions by a composite score derived from real network usage.
A new publishable embed token lets the pricing snippet call the API straight from your storefront, restricted to the sites you list, with your secret key staying on your server.
A Model Context Protocol server lets AI agents compute prices through Last Price with a single tool call, no custom integration required.
A new Last Price mobile app lets you price any photo, document, audio clip, or product link from your phone.
The Last Price mobile app has a sharper, high-contrast look and a simplified pricing screen that puts one Add action front and center.
The Last Price mobile app now accepts photos, screenshots, and links shared directly from the OS share sheet, with reliable attachment handling and precise cost display.
The security criteria governing multi-tenant access were strengthened, and the documentation for how tenant API keys are stored was corrected to reflect actual behaviour.
41 schema definitions that had been dropped from the OpenAPI spec are now back in place, keeping generated clients and agent integrations accurate.
The published OpenAPI spec now documents the correct authentication header, maps Elo routes accurately, and gives the function-test operation its own request schema.
Clicking an order in the list opens the order instead of a missing page, and the order now shows its line items.
The integration guide, deployment reference, and network design documents were rewritten to match how the API actually behaves, with accurate request examples, correct credentials, and honest notes about what is and is not yet live.
Experiments that could never run are rejected when you create them, and the traffic split you configure is now applied.
Usage and cost views now include single price calls, which were previously counted only for batch requests.
The task cost estimator now returns live per-country estimates from the AI model instead of sample data.
A new GET /api/me reports the workspace, auth method, and scopes behind any credential, and the dashboard now shows the workspace ID instead of asking you to type it. The function test endpoint no longer 404s on functions you just created.
The marketplace now lists pricing models for media businesses, with a dedicated Media category covering streaming, ad inventory, stock licensing, and generative render workloads.
A new "Load sample data" button on the Commerce dashboard sets up a demo store, competitor prices, and a pricing policy, then runs a cycle so you can see and approve a real price proposal without connecting a store.
Several pages that previously showed empty tables or stale data under error conditions now load correctly and surface actionable status messages.
The costs, usage, invoice detail, and payment detail pages were previously showing placeholder or invented data. They now read from and write to the live APIs, so every number and action reflects your actual account.
Creating a payment now saves the correct date and shows it in the payments list. The invoice detail Edit link and the order edit page no longer fail with an undefined record ID.
A new public /price-check page lets anyone query live multi-modal pricing without signing in. Results stream through the public pipelines endpoint with a 2 MB body cap.
A scheduled job now builds and publishes a structured pricing dataset, keeping downstream consumers in sync with the latest provider prices.
The pricing models page now loads live data from a new dedicated API endpoint, replacing the previous mock state.
The usage signals page is now connected to a live API, replacing the previous mock data. The API validates date and limit parameters and returns clear errors on bad input.
Webhook endpoints you create are now saved to your account, survive page reloads, and support a signed test ping with a per-delivery log.
The API now accepts first-party session tokens alongside API keys, and the x-tenant-id header is validated against your workspace membership before taking effect.
PriceRouter GET and POST endpoints are now fully documented in the OpenAPI spec with typed schemas for capability, request, and response. A dedicated feature page makes the multi-vendor router discoverable from the marketing site.
Last Price now speaks UCP natively. AI agents discover capabilities at /.well-known/ucp, create checkout sessions, get optimized pricing quotes, and complete purchases, all through one standard protocol.
Commerce connector credentials (Shopify, Amazon SP-API, and friends) are encrypted at rest with AES-256-GCM envelope encryption before they touch the database. Plaintext fallback has been removed.
We rebuilt the account flow from scratch. Sign in and sign up have their own pages, sessions refresh automatically as you browse, and Google sign-in lands you in the dashboard in one click.
You can now browse parts of the dashboard without an account so prospects and AI agents can see what Last Price actually looks like before signing up. Anything that writes still requires auth.
Every API route now derives the caller from the verified bearer token. The legacy x-user-id request header fallback has been removed across components, bucket-key paths, and the OpenAPI spec.
Connect a storefront, ingest price observations, match them to your catalog, decide on a price, route the change through approvals, and execute back to the store. All seven steps now run from one cycle.
New LastPriceLogo squircle, refreshed color tokens, and a single shared wordmark component so the dashboard, marketing site, and auth flows finally look like one product.
One OpenAPI document now describes every endpoint we expose: PriceRouter, Rosetta, Commerce, Signals, and the inference router. Generate a client in your language, or hand it to an agent.
Every call to PriceRouter can now explain itself, and you can compare strategies side by side before you commit. Agents get a downloadable tool manifest so they can wire up PriceRouter with no human in the loop.
PriceRouter exposes every dynamic-pricing vendor we host behind one API. The capability descriptor lists providers and models so AI agents can pick the right one without hand-coded routing.
Privacy Policy and Terms of Use have been rewritten in plain language. The old /tos URL now permanently redirects to /terms, and a new /legal hub links to both.
Loading a workspace no longer scans the full team list to confirm you are a member. The membership check is now a single indexed row fetch, and the team page loads in one round trip.
A browser-side tool at /product/webhooks/verify that confirms a webhook payload was actually sent by us. Paste the signature header and body, get a green check or a precise reason it failed.
Bring your team into Last Price. Each workspace has owners, admins, and members; every API call and dashboard action is scoped to the workspace you are signed into.
We added seven built-in functions across Rosetta, Elo, and Jale, bringing the catalog to fifteen pricing primitives you can call by name with no setup.
One call to POST /api/rosetta/ingest takes a merchant payload, proposes a canonical mapping, validates it against the catalog schemas, and stages the result for review.
The dashboard now opens in dark mode unless your OS asks for light, and every page works on a phone. Tables collapse to cards, dialogs stop clipping, and tap targets meet the 44 by 44 minimum.
Rosetta gives you one canonical schema for offers, products, observations, matches, and actions. The new Mapping Studio lets you draw the mapping in the UI; the new translate API runs it at scale.
Workspace API keys are now AES-256-GCM encrypted before they touch the database. Dashboard sessions also stop surviving a closed tab.
Drop a one-line snippet on your storefront to feed Last Price live context, and run pricing variants through a new Experiments page with significance testing built in.