Kalcend vs Pickaxe (2026)

Stands at F, with the count corrected and the breadth verified rather than summarised. The July basis said over 40 tools; the vendor states 43 one-click integrations, and naming the exact number is the kind of precision worth preserving. BREADTH ACROSS CLASSES IS MET PROPERLY, which is what this axis measures and what a large single-ecosystem catalogue fails. The named classes are CRM with HubSpot, Salesforce, Zoho and Capsule, helpdesk with Intercom, Zendesk and Zoho Desk, and Google Workspace across Gmail, Calendar and Drive, alongside payments through Stripe and documents through Notion. THE CAPABILITIES ARE THE UNDERRATED PART and none was in the July basis. Agents READ AND WRITE FILES, RUN SQL OVER CSVS AND SHEETS, QUERY AND WRITE TO YOUR DATABASE, and RUN SHELL COMMANDS OVER SSH. Direct database write access and SSH shell execution are a materially wider action surface than a connector list implies, and for a product sold to non-technical buyers they are a surprising amount of power to hand an agent. CUSTOM INTEGRATIONS CLOSE THE CEILING. The Builder designs, builds and deploys custom integrations, webhook receivers and internal APIs from the chat flow, so coverage is not bounded by the 43. Signed webhooks indicate the inbound path is real. Browser automation reaches what none of the above can and is credited on Comp rather than counted twice here. One limit: no integration catalogue page was reached, so the 43 rest on the features page summary and the classes on its named examples.

Stands at F, now anchored on the vendor's own browsable Actions directory rather than on a summary sentence. BREADTH ACROSS CLASSES IS DOCUMENTED AND VERIFIABLE. The Actions directory at pickaxe.co/actions is an enumerated catalogue rather than a claimed count, spanning CRM with HubSpot, Salesforce and Attio, communication with Gmail and Slack, productivity with Notion, Google Calendar and Google Sheets, developer tooling with GitHub, and generation with DALL-E, alongside Composio for onward tool authentication. Those are genuinely different integration classes, which is what separates this from a deep catalogue inside one ecosystem. TOOL CALLING IS REAL AND BIDIRECTIONAL. The vendor's framing is that Actions let YOU AND YOUR USERS TAKE ACTIONS AND TRIGGER WORKFLOWS WITH THE TOOLS IN REAL TIME, so the agent both reads and writes, and the end user of a deployed agent inherits that reach rather than only the creator. THE MCP DIRECTORY IS THE PART THAT SCALES IT. Pickaxe publishes MCP servers as Actions, so any MCP-exposed service becomes agent-callable without Pickaxe building a connector for it. That is the inbound direction, an agent calling out through MCP, and it is credited here rather than on Ext, where the outbound direction sits. Custom Actions plus Zapier and Make webhooks close the ceiling for anything not in the catalogue. ONE COMMERCIAL LIMIT WORTH RECORDING because it affects what a buyer actually gets: the Gold tier caps Actions at three per agent, with the cap removed on Pro and Business. A capability gated by plan is still the capability, but the entry tier is materially narrower than the 500-plus figure suggests.

Stands at F, re-based onto the multi-agent and escalation mechanics rather than the general claim that the Builder assembles an agent. THE MULTI-AGENT CLAIM IS SPECIFIC AND WAS NOT IN THE JULY BASIS: AGENTS TRIGGER AND NOTIFY EACH OTHER, AND ESCALATE TO A PERSON WHEN NEEDED. Agent-to-agent triggering with a human as a third participant is genuine multi-party coordination, and stating it in one sentence alongside human escalation shows the routing model treats people and agents as peers in the same flow. MULTI-STEP EXECUTION IS DOCUMENTED THROUGH THE WORK ITSELF rather than through a graph. A single hire qualifies leads, schedules meetings, updates the CRM and triggers downstream workflows, and the underlying action set spans file reads and writes, SQL over spreadsheets, database writes, SSH commands and browser navigation. Sequencing across that many action types is orchestration whether or not a canvas exists. THE ARCHITECTURAL POINT IS THAT DURABILITY IS PART OF ORCHESTRATION HERE. Long-running stateful agents on messaging infrastructure that handles delivery, retries and ordering means a multi-step task survives a customer replying an hour later, which is the failure mode that breaks naive agent loops in conversational products. WHAT IS DELIBERATELY ABSENT IS CONTROL-FLOW VOCABULARY, and the vendor says so: THERE IS NO FLOW BUILDER, NO SCRIPTING, NO SEPARATE TECHNICAL WORKFLOW FOR POWER USERS. No branching, looping or conditional construct is named because the design intends the model to decide the path. That is a legitimate architecture rather than a gap, and it is the axis' harder case: the orchestration is real but nothing lets a customer inspect or constrain its shape. Confidence is medium for that reason.

Stands at F, with the multi-agent mechanism named more precisely than the July basis managed. SUB-AGENT ROUTING IS THE LOAD-BEARING FACT: an agent routes to specialised sub-agents in a WATERFALL configuration for multi-step tasks. A waterfall is an ordered fallback chain rather than a parallel fan-out, which is a modest but real orchestration primitive and an honest description of what a no-code builder for solo creators would actually ship. MULTI-STEP EXECUTION IS DOCUMENTED THROUGH THE WORK. Post-submission workflows fire after a conversational form completes, Actions chain into connected tools, and the action set spans PDF generation, structured data collection through conversational forms, and code execution. Sequencing across generation, collection and external writes is orchestration whether or not a canvas exists. THE ABSENT PRIMITIVE IS THE HONEST LIMIT AND IT IS THE SAME ONE AS ELSEWHERE IN THIS LANE. No branching, looping, conditional or parallel construct is named on any page reached, and there is no visual flow canvas. Control flow is the model's decision plus a waterfall, which suits the product's audience and would not suit a process that must run the same way every time. CONFIDENCE IS MEDIUM RATHER THAN HIGH for that reason: the grade rests on documented multi-agent routing and step sequencing, which are present, rather than on an inspectable orchestration surface, which is not. This is the cell most likely to read differently to a buyer coming from n8n or FlowX than the letter suggests, and worth a note on any comparison page pairing them. Code execution is recorded here as an action type and is deliberately refused on Comp, per the standing convention that executing code is not operating software.

Stands at F and is the best-evidenced cell on the record, because channel coverage is the product's headline claim and it is stated identically on the home page, the features page and the privacy policy. FIVE CHANNELS RUN SIMULTANEOUSLY OUT OF THE BOX: WhatsApp Business API, Slack, inbound voice calls, web chat, and email. Breadth is not the interesting part; SIMULTANEOUSLY is. One configured agent is reachable on all five at once rather than being deployed per channel, which is why the shared cross-channel memory credited on Mem is necessary rather than decorative. VOICE IS THE ENTRY THAT SEPARATES THIS FROM THE LANE. Inbound phone handling is a materially harder channel than text and most agent builders do not attempt it, so its presence alongside four text channels is a real coverage claim rather than a list padded to look wide. THE EVENT CLASS IS COVERED SEPARATELY AND IN TWO FORMS. Agents trigger workflows in connected tools, and, more interestingly, AGENTS TRIGGER AND NOTIFY EACH OTHER, so work can reach an agent from another agent. The privacy policy corroborates by logging workflow and webhook activity as a distinct data category, and webhook receivers built by the Builder give an inbound programmatic route. WHAT IS NOT DOCUMENTED IS THE SCHEDULE CLASS. No timer, cron or recurrence appears on any page reached. For a product built around inbound customer conversation that is a coherent omission rather than a gap, and the channel and event classes together carry the grade comfortably without it. Campaign sending is documented in the privacy policy as an outbound activity, which implies scheduling but is not credited from an implication.

Stands at F and is the best-evidenced channel cell reviewed this session, because coverage is stated per channel with the delivery mechanics named rather than as a list. SIX ROUTES ARE DOCUMENTED INDIVIDUALLY: website embeds in inline, floating, popup and iframe forms; branded portals on custom domains with user accounts and access groups; a Completions API; email; Slack; and WhatsApp, with third-party review adding Telegram and Discord. THE EMAIL DETAIL IS THE ONE WORTH CARRYING. The vendor documents running AN AGENT ON A REAL EMAIL ADDRESS WITH THREADED MEMORY AND ATTACHMENTS. Threading and attachments are the two things that make email genuinely usable as an agent channel rather than a notification pipe, and most platforms claiming email coverage do neither. THE EVENT CLASS IS COVERED SEPARATELY IN BOTH DIRECTIONS. Actions fire workflows in connected tools in real time, and outbound webhooks notify Zapier, Make and other endpoints. The Completions API gives an inbound programmatic route, and authenticated embeds handle session handoff from a host application. WHAT IS ABSENT IS THE SCHEDULE CLASS. No timer, cron or recurrence appears on any page reached. For a product whose agents are summoned by an end user rather than run unattended that is a coherent omission, and channel plus event coverage carries the grade comfortably without it. A property specific to this vendor: because portals carry accounts and access groups, a channel here also carries identity, so the same agent recognises a returning user across visits. That is what makes the persistent per-user memory credited on Mem usable rather than theoretical.

Stands at F, and the privacy policy turns what was a product-page claim into a described mechanism. THE RETRIEVAL STACK IS ENUMERATED IN THE LEGAL PAGE: knowledge base content is UPLOADED DOCUMENTS, URLS, TEXT CHUNKS, EMBEDDINGS. Counsel listing chunks and embeddings as categories of stored customer data is an unusually concrete confirmation that a real vector index exists, because those are implementation artefacts rather than marketing nouns. RETRIEVAL IS STATED AS THE MECHANISM: the vendor GENERATES ANSWERS USING CUSTOMER CONTENT VIA RETRIEVAL, with responses stored for audit and team review and not used to train third-party models. That last clause matters to the buyer this product serves and is the kind of commitment that appears in a DPA rather than on a landing page. THE PERSISTENCE LINE IS CLEARED. The knowledge base is a maintained, addressable object generated once from the customer's material and queried thereafter, not context assembled per request. The Builder GENERATES A CUSTOM KNOWLEDGE BASE FROM YOUR DOCS as a build step, which is the structure this axis asks for. SOURCE BREADTH IS WIDER THAN DOCUMENTS ALONE and this is where it separates from peers: answers come from PDFs, docs and spreadsheets, and agents additionally run SQL over CSVs and sheets and query the customer's database directly. Grounding on a live database rather than only on an uploaded corpus means the agent reads current state, which for support and sales work is the difference between a correct answer and a stale one. One limit: no chunking strategy, embedding model, re-indexing or freshness behaviour is documented.

Stands at F and clears the persistence line comfortably, which is the discriminator this axis turns on. AN INTEGRATED VECTOR DATABASE IS THE MAINTAINED STRUCTURE. The knowledge base is a first-class object the creator builds once and the agent queries thereafter, not context assembled per request, and it is shared at workspace level so several agents ground on the same corpus. SOURCE BREADTH IS THE STRONGEST PART: more than 14 content types including PDFs, websites, Notion, Google Drive and video. Video ingestion is uncommon in this lane and matters for the buyer Pickaxe serves, since coaches and consultants hold much of their expertise in recorded material rather than documents. THE FRESHNESS MECHANISM IS WHAT MOST PEERS LACK. AUTOMATIC DAILY SYNCS mean a knowledge base connected to Notion or Drive tracks the source rather than freezing at upload. The usual failure of a document-index product is a corpus that silently goes stale, and a documented sync cadence is the direct answer to it. PER-FILE CONTEXT INSTRUCTIONS ARE THE DETAIL WORTH CARRYING. A creator annotates individual documents with guidance on how the agent should use them, which is retrieval steering at the document level rather than one global prompt. For a knowledge base assembled from mixed material, a price list, a policy, a transcript, telling the agent what each is for is a meaningful quality lever. The CLI syncs knowledge bases programmatically, so the corpus is maintainable from code as well as the web app. What is not documented is chunking strategy, embedding model, retrieval configuration or citation behaviour in responses.

Stands at F under the state half of the axis, and the features page supplies the term the July basis lacked: LONG-RUNNING STATEFUL AGENTS, named as one of the three things the platform is built on. STATE IS ARCHITECTURAL HERE RATHER THAN A FEATURE. The vendor draws the contrast itself: when a customer messages an agent, IT DOESN'T POKE A PROCESS IN A CONTAINER, it flows through messaging infrastructure handling delivery, retries and ordering. An agent that survives between messages, holds ordering, and resumes is a persistent process rather than a stateless function invoked per turn. THE CROSS-CHANNEL PROPERTY IS THE DISTINCTIVE PART. Full conversation memory is shared across all channels, so the same hire keeps context whether the customer arrives on WhatsApp, phone, Slack or email. That is memory keyed to a person rather than to a session, which is a stronger claim than session persistence: the agent must resolve identity across channels to apply it, and it is exactly the property a support buyer notices when a customer switches from chat to a phone call mid-issue. Under the 31 August ruling this clears Full on the state limb: state persists across sessions and is addressable by the agent, and accumulation is one route to Full but never the required one. Partial is session-only state, which this explicitly is not. The escalation path corroborates it, since handing a conversation to a person WITH THE FULL THREAD ATTACHED requires the thread to have been retained and retrievable. One limit: no retention period, memory inspection surface, or means of editing or clearing what an agent remembers is documented.

P>F, and it clears both limbs of the 31 August ruling rather than only the state one, which is unusual. THE STATE LIMB IS STATED IN THE VENDOR'S OWN WORDS. The features page says the platform gives agents PERSISTENT MEMORY ACROSS SESSIONS. That is the ruled Full condition verbatim, and the July basis missed it by grading from the Configure tab's context-window token allocation, which is a within-session concern and correctly reads as Partial on its own. THE ACCUMULATION LIMB IS MET TOO AND IS THE BETTER FINDING. A USER MEMORIES tab lets the operator CONFIGURE TRAITS THAT AGENTS WILL COLLECT AND REMEMBER ABOUT USERS ON A USER-BY-USER BASIS, naming a user's name, role and preferences as examples. Memory that the agent gathers over time, keyed per end user and shaped by operator-declared traits, is the accumulation route, and configurable extraction targets are a more deliberate design than a generic memory store. THIS IS ALSO ADDRESSABLE, WHICH THE RULING REQUIRES. Memory is exposed as an API resource rather than an internal effect: a community MCP server built against the public API calls memory_list and memory_get_user to read per-user memory. That is third-party code and carries nothing on its own, but it demonstrates the memory objects are real, named and reachable through the documented API. WHY THIS MATTERS FOR THE PRODUCT rather than being a checkbox: Pickaxe agents are sold to end users through branded portals with accounts, so per-user memory is what makes a purchased agent feel like a returning assistant rather than a stateless chatbot. The capability is load-bearing for the monetisation model. What is not documented is retention period, or how an operator or end user inspects, edits or deletes stored memories, beyond the fact that the API can list them.

P>F. The July basis held Partial because a runtime approval gate was not detailed, and cited the sandboxed playground instead. The playground is build-time testing and belongs on Eval; the runtime gate exists and is described plainly. THE MECHANISM, VERBATIM: AGENTS PAUSE FOR APPROVALS, SIGN-IN, AND OTP, AND HAND CONTROL TO A PERSON FOR ANYTHING SENSITIVE. WHEN A REQUEST NEEDS JUDGMENT, IT ESCALATES TO YOUR TEAM ON SLACK OR A SHARED INBOX, WITH THE FULL THREAD ATTACHED. That is a pause, a handoff, a destination and context transfer, which is every element this axis asks for. THE DETAIL THAT MAKES IT CREDIBLE RATHER THAN ASPIRATIONAL IS THE FULL THREAD ATTACHED. Escalation without context is a ticket; escalation with the conversation attached is a handoff a person can actually act on, and it is the same property that distinguishes a real support handoff from a dropped one. HANDING CONTROL BACK IS THE STRONGER FORM. Most approval mechanisms in this lane pause and wait for a yes or no. This one transfers control of the conversation to a human, who then works it in Slack or a shared inbox. For a support and sales product running on WhatsApp and voice, that is the correct shape, because the alternative to the agent continuing is a person continuing, not the work stopping. SIGN-IN AND OTP AS NAMED PAUSE CONDITIONS tie this directly to the browser automation credited on Comp: the agent drives a site until it hits authentication, then hands the keyboard over. That is a specific, unfaked design decision rather than a governance sentence. What is not documented is configurability: no policy language, threshold, or per-action approval rule was reached, so which actions count as SENSITIVE appears to be the platform's judgment rather than the customer's. That is the natural check at lane close.

Stands at P, with the guardrail half strengthened by a mechanism the July basis did not have and the human half still absent. THE ADDITION IS SANDBOXED AGENT EXECUTION THROUGH THE OPENCLAW ENGINE, offered to regulated-industry customers. Execution isolation is a genuine containment guardrail and a stronger one than configuration limits, because it constrains what an agent can do at runtime rather than what it was told to do. It is named as an engine rather than a policy, which suggests a real boundary. AROUND IT SIT THE CONSTRAINTS THE JULY BASIS FOUND: access groups across public, member and invite-only, per-user usage budgets, role-based access controls and granular permissions, and per-file context instructions steering how documents are used. Those are real limits on who reaches an agent, how much they can consume, and what it draws on. WHAT IS STILL MISSING IS THE HUMAN HALF, and the July basis was right to say so. Nothing pauses an agent mid-run pending a person's approval, no pending-action queue exists, and no escalation or handoff path to a human is documented on any page reached across two passes. For a platform whose Actions write to HubSpot and Salesforce and whose agents are sold to third-party end users, the absence of an approval step before an external write is the notable gap. THE COMMERCIAL SHAPE EXPLAINS IT WITHOUT EXCUSING IT. Pickaxe agents are sold BY a creator TO end users, so the creator is not present during a run and there is nobody to approve; oversight is necessarily front-loaded into configuration and after-the-fact review. That is a coherent design and Partial describes it accurately. Prompt Injectors appear in the API documentation and may bear on runtime constraint; the page was not reached and is worth checking at lane close.

Stands at P, but on a completely different and much better basis, and confidence rises from low to medium. WHAT CAME OUT: the July basis rested Partial on enterprises running custom models in their own infrastructure. That claim is aggregator-only, excluded, and it is a deployment or model fact rather than a security control in any case. Removing it left the cell with nothing, and I expected to drop it to None. WHAT REPLACED IT IS A GENUINELY SUBSTANTIVE PRIVACY PROGRAMME, which is the silna-health shape: real engagement with obligations, no certification. The privacy policy performs an actual controller and processor analysis, naming the vendor as data controller for account data and processor for end-user data, and states a DATA PROCESSING ADDENDUM APPLIES to data processed on a customer's behalf. Integration credentials and tokens are STORED ENCRYPTED. Processing occurs ONLY ON THE CUSTOMER'S DOCUMENTED INSTRUCTIONS. Rate-limiting, spam detection and incident response are named. TWO DETAILS RAISE IT ABOVE BOILERPLATE. The vendor discloses its model subprocessors by name, OpenAI and Anthropic, and states it uses their APIs WHERE API DATA IS NOT USED TO TRAIN THEIR MODELS. And it excludes special categories of data, health, biometric and financial account numbers, unless a written DPA permits them. A vendor telling prospects what not to put in the product is making a scoping commitment against its own commercial interest, which boilerplate does not do. WHY IT IS NOT FULL: the F bar is a conjunction requiring an attestation or certification plus a named customer-facing control. No SOC 2, ISO, trust page, SSO, role-based access control or retention configuration was reached across two passes. This is the half of the conjunction that is missing, and for a grant-stage company months from launch that is expected rather than alarming.

Stands at F, and confidence rises because the July basis rested on one sentence from the features page while two dedicated security pages exist that it never reached. THE CONJUNCTION IS MET ON BOTH HALVES, WHICH IS WHAT THE 30 AUGUST BAR ASKS. Attestation: the vendor states it has been INDEPENDENTLY EXAMINED FOR SECURITY, AVAILABILITY, AND CONFIDENTIALITY CONTROLS, that SOC 2 REPORTS ARE AVAILABLE TO ENTERPRISE CUSTOMERS THROUGH OUR TRUST CENTER, and claims GDPR and CCPA compliance. Named customer-facing controls: enterprise single sign-on, role-based access controls, granular permissions, comprehensive audit logging, and configurable data retention. Four named controls where the bar requires one. A REPORT THAT EXISTS AND CAN BE REQUESTED IS MATERIALLY ABOVE THE HEDGE LADDER, whose rungs describe assertions with no report behind them. This is a completed examination with three trust services criteria named and a Trust Center as the delivery route. THE ONE HEDGE WORTH FLAGGING, AND IT IS A DELIBERATE WORD CHOICE: the vendor says EXAMINED throughout and never says TYPE II. Type is unspecified on every page reached, which on the ladder alone would sit at Partial. I have graded Full on the conjunction, since the controls half is documented in detail and a report is obtainable, but a procurement team should establish the type and period through the Trust Center before relying on it. That is the single thing to verify on this cell. TWO SUBSTANTIVE COMMITMENTS BEYOND THE CHECKLIST: knowledge base content and conversation data are NEVER USED TO TRAIN AI MODELS, and regulated-industry customers are offered sandboxed agent execution through the OpenClaw engine. Per standing practice the deployment position is graded on Dep, where it is None, and is not borrowed here.

Stands at P, re-based, with the sandbox playground removed because it was one fact working three axes. WHAT CAME OUT AND WHY. The July record cited the sandboxed playground on Human Oversight, on Observability AND on Testing. A build-time simulation environment is a testing surface; it is not oversight of a running agent and it is not a record of what one did. It now sits only on Eval, and this cell and HITL were re-decided on their own evidence, which moved HITL up and left this one where it was. WHAT ACTUALLY SUPPORTS PARTIAL IS IN THE PRIVACY POLICY, which enumerates what the platform records: AGENT AND AUTOMATION ACTIVITY, TASKS AGENTS PERFORM, INCLUDING WEB PAGES AND DATA ACCESSED VIA BROWSER AUTOMATION AND CONNECTED INTEGRATIONS, PLUS WORKFLOW AND WEBHOOK LOGS. That is an activity record covering the action surface, and it is corroborated by responses being STORED FOR AUDIT AND TEAM REVIEW, which states the customer's team reads them. Recording which web pages an agent visited is a meaningful audit artefact for a product whose agents drive browsers, and it is more specific than most vendors in this lane document. WHY IT DOES NOT REACH FULL, and this is the reporting-versus-auditing line. What is documented is what happened, tasks performed, pages accessed, workflows fired. Nothing documents why: no reasoning trace, no per-step decision record, no tool-call detail, no run inspection surface. The escalation path attaching a full thread is conversation history rather than an execution trace. No retention period, export path or dashboard was reached. Confidence rises from low to medium because the cell now rests on an enumerated legal-page description of what is logged rather than on a build-time playground.

P>F. The July basis withheld Full because detailed run tracing or audit logging was NOT DOCUMENTED AS FIRST CLASS. It is documented as first class on two pages the build did not reach. THE AUDIT HALF IS EXPLICIT AND COMPLIANCE-FACING: COMPREHENSIVE AUDIT LOGGING TRACKS EVERY INTERACTION, GIVING YOUR COMPLIANCE TEAM THE DOCUMENTATION THEY NEED FOR PROCUREMENT REVIEWS AND REGULATORY AUDITS, with the security page adding that operators track every interaction, monitor usage patterns and know exactly how the AI is being used, who is using it and when. THE INSPECTION HALF IS THE PART THAT DECIDES IT, and it comes from the user manual. An ACTIVITY tab shows usage across agents, EACH CONVERSATION HAS A HIGH-LEVEL AI-GENERATED SUMMARY, AND YOU CAN CLICK IN TO SEE THE FULL CONVERSATION IF SOMETHING NEEDS INVESTIGATING. Summary-then-drill-in is a genuine investigation workflow rather than a log dump: an operator with hundreds of conversations can find the one that went wrong, which is what auditability means in practice. FOR A CONVERSATIONAL AGENT THE TRANSCRIPT CARRIES MOST OF THE WHY. Unlike a background automation where reasoning is invisible unless traced, an agent that talks leaves its reasoning in the conversation, so full-transcript retrieval plus per-user attribution reconstructs what happened and largely why. THE HONEST GAP, and the reason confidence is medium: Action execution is not separately recorded on any page reached. For a platform with 500-plus Actions writing to HubSpot and Salesforce, a per-Action call record with parameters and results is the missing artefact, and a compliance team could see that a conversation occurred without seeing which CRM record it changed. Usage tracking, per-user credit accounting and analytics sit alongside. Retention period and export are undocumented.

P>N. The Partial rested on a single fact, that enterprises can run custom models inside their own infrastructure, and the July basis cited THE SMOL LAUNCH LISTING for it. That is a launch-directory listing, which section 7 excludes as evidence outright. I CHECKED WHETHER THE CLAIM APPEARS FIRST-PARTY AND IT DOES NOT. The phrase recurs across Smol Launch, Crunchbase and other directory profiles, which is what a single submitted blurb propagating through aggregators looks like, and none of those is evidence. It appears on no page of kalcend.ai reached across two passes. WHAT THE VENDOR'S OWN PAGES SAY IS THE OPPOSITE SHAPE. The privacy policy describes a SaaS platform operated by an Indian company, acting as data controller for account data and processor for end-user data, running on the vendor's own infrastructure with named subprocessors. The company's Kerala Startup Mission profile describes agents built, deployed and run IN A MANAGED CLOUD. Nothing documents self-hosting, a virtual private cloud option, region selection or customer-managed keys. UNDER SECTION 7 SINGLE-REGION SAAS WITH NO SELECTION IS NONE, and that is what this is. The grade asserts the capability is not documented as first-class, not that the vendor is careless: the privacy policy is unusually substantive about where data goes. ONE FACT MOVED RATHER THAN DELETED, because it was doing work on three axes at once. The custom-models claim appeared on Security, Deployment and Model Flexibility simultaneously. Even if it were first-party it would be a model fact and would be spent once, on Model. Spending one unverified sentence three times is how a thin record reaches a composite of nine. If a lane-close pass finds an enterprise or self-hosting page on kalcend.ai, this returns to Partial or better on that page.

P>N, and the correction is a category one: the July basis graded distribution surfaces as deployment. WHAT IT CITED was rich multi-surface deployment through website embeds, branded portals with custom domains, API, email, WhatsApp and Slack. Every one of those is a channel by which an end user REACHES the agent, and every one is already credited on Trig, where they belong and where they earn Full. Where a chat widget renders says nothing about where the software runs or where the data sits, which is what this axis measures. The same basis then conceded the actual answer: THE PLATFORM IS CLOUD ONLY AND NO SELF HOSTED OR DATA RESIDENCY OPTIONS ARE DOCUMENTED. THIS PASS CONFIRMS THAT CONCESSION AGAINST THE VENDOR'S OWN SECURITY PAGES, which are where a residency option would be advertised if one existed. They are unusually detailed, covering encryption at rest and in transit, data ownership, retention configuration, a Trust Center and custom SLAs, and they name no region selection, no virtual private cloud, no on-premises option and no data localisation commitment. A vendor documenting this much and omitting residency is not hiding it. UNDER SECTION 7 SINGLE-REGION SAAS WITH NO SELECTION IS NONE, and that is the correct reading. Grading Partial here would credit the platform twice for the same channel list. TWO ADJACENT FACTS RECORDED AND DELIBERATELY NOT CREDITED. Retention configuration is a control and is credited on Sec. SANDBOXED AGENT EXECUTION THROUGH THE OPENCLAW ENGINE, offered to regulated industries, is runtime isolation rather than deployment location, and is credited as a guardrail on HITL. ADVANCED DATA PRIVACY OPTIONS are offered to enterprise customers without being specified. If those turn out to include residency, this cell moves; it is the natural question for a lane-close pass through the Trust Center.

P>N under the 31 August Pack bar, which asks whether the customer receives packaged assets ready to adopt by any route. THE JULY BASIS CITED TEMPLATES AND ROLE PRESETS, an inbound phone agent that books appointments and a WhatsApp triage agent. Those examples appear in launch-directory and company-profile listings, which are excluded as evidence, and on no page of kalcend.ai reached across two passes. Read carefully, the listings describe what you can ASK FOR rather than what ships: a receptionist that answers calls, an assistant that handles emails, a bot that manages WhatsApp are example prompts, not catalogue entries. WHAT THE PRODUCT ACTUALLY DOES IS THE REASON THIS IS NONE RATHER THAN UNRETRIEVED, and it is a genuinely interesting negative. The Builder LEARNS YOUR BUSINESS, GENERATES A CUSTOM KNOWLEDGE BASE FROM YOUR DOCS, WRITES REAL INTEGRATIONS, and the vendor states there is no flow builder and no separate technical path. Generation from a prompt is this platform's deliberate substitute for a template library. The vendor's own positioning is that it hires an employee rather than assembling one from building blocks. SO THE ABSENCE IS ARCHITECTURAL, NOT A GAP. A catalogue would be redundant in a product whose premise is that you describe the role and it builds the specific agent. Recording None here is accurate about what a buyer receives and should not be read as a deficiency in the way an empty Security cell would be. WORTH CARRYING TO COMPARISON PAGES: against lindy, relevance-ai and stackai, which this record names as peers, Kalcend trades a browsable starting library for generation. That is a real product difference and it is exactly what this axis is for. The 43 one-click integrations are connectors and are credited on Int.

Stands at P, with one fact removed for working two axes and one new lead recorded that could move it. WHAT CAME OUT: the prebuilt Actions library. Actions are integrations and are credited on Int, where they carry Full on an enumerated public directory. Counting the same library here would credit one catalogue twice, and it is not what this axis measures anyway: a connector is a capability the customer wires in, not a packaged agent they adopt. WHAT REMAINS IS GENUINELY MIXED. A gallery of use-case examples shows what others have built, which is illustration rather than an adoptable asset. The AUTOMATIC AI BUILDER generates an agent from a description, which is the kalcend pattern seen earlier today: generation substituting for a catalogue. Neither is a packaged asset ready to adopt. THE LEAD THAT COULD MOVE THIS CELL, recorded rather than graded: site furniture on the Actions pages reads GET STARTED FASTER WITH PRE-BUILT TEMPLATES, CHOOSE FROM OUR LIBRARY OF READY-TO-USE AI TOOLS AND CUSTOMIZE THEM FOR YOUR NEEDS. Under the 31 August bar a library of ready-to-use tools would clear Full, and the phrasing is a direct first-party claim. I have not graded from it because I reached the sentence and not the library, and I am not willing to grade a catalogue I have not seen after removing another catalogue for being counted twice. Locating that template library is the single check for this cell at lane close. WORTH NOTING FOR THE COMMERCIAL MODEL: Pickaxe's premise is that creators PACKAGE AND SELL agents, so the marketplace it operates is the creators' storefronts rather than a vendor-supplied pack library. The absence is coherent with the business rather than an oversight.

Stands at P, but the fact carrying it is replaced, because the July basis rested on an aggregator claim and inferred flexibility from it. WHAT CAME OUT: enterprises can run custom models inside their own infrastructure. That sentence appears only in launch-directory and company-profile listings, never on kalcend.ai across two passes, and section 7 excludes such listings outright. It was also being spent three times, on Security, Deployment and here. Inferring model flexibility from an unverified deployment claim is two errors stacked. WHAT REPLACES IT IS DISCLOSURE, NOT SELECTION, and it comes from the privacy policy: the vendor names OPENAI AND ANTHROPIC as its model subprocessors and states it uses their APIs where data is not used to train their models. Two named providers, chosen by the vendor. UNDER THE STANDING CONVENTION THAT IS PARTIAL RATHER THAN NONE. Vendor-internal use of more than one provider without customer selection sits at Partial; a single named provider would be the best-evidenced None. Two providers named in a legal page is a first-party statement that more than one model powers the product, which clears the Partial floor without approaching customer choice. WHAT IS ABSENT IS THE WHOLE OF WHAT THIS AXIS MEASURES: no model picker, no bring-your-own-key, no endpoint configuration, no per-agent model setting and no documented routing logic. Nothing lets a buyer choose what powers their agent. THAT IS COHERENT FOR THE PRODUCT and worth saying rather than treating as a defect. A platform whose promise is that a non-technical person describes a role in plain English and receives a working employee is one where model selection is a decision the vendor makes so the customer does not have to. The disclosure is honest; the choice simply is not offered.

Stands at F and is one of the cleanest Model cells in the lane, because the customer's choice is both real and cheap to revise. MORE THAN 50 MODELS ACROSS OPENAI AND ANTHROPIC with ONE-CLICK SWITCHING. The switching cost is what earns Full rather than the count: a choice made at build time and then frozen is weaker than one a creator revisits when a cheaper or better model ships, and one click is as low as that cost goes. THE COST DIMENSION IS THE PART WORTH CARRYING and it is unusual in this lane. Live model rates are shown in the product, and a cost estimator lets a creator compare price against speed and quality before committing. Pickaxe's economics make this load-bearing rather than decorative: creators resell agents and end-user usage draws down the creator's credits at one dollar of credit per dollar of model cost, so model choice is directly the creator's margin. A platform that shows live rates at the point of selection is helping its customers price their own products. TWO PROVIDERS RATHER THAN A LONG PROVIDER LIST is the honest limit. Google, Meta, Mistral and open-weight models are not documented, so a customer wanting a specific non-OpenAI, non-Anthropic model cannot have it, and there is no bring-your-own-key or custom endpoint path documented. Within those two families the depth is real. No automatic routing between models at runtime is documented, which this axis does not require; it measures customer choice, which is amply met. Per section 7 the Actions surface and the API run the other direction and are credited on Int and Ext rather than counted here.

P>F under Mike's 30 August bar, which asks whether the platform is callable from outside through a documented API or SDK. THE JULY BASIS WAS CAREFULLY WRONG. It noted, correctly, that the Builder creating custom tools, webhooks and internal APIs is the customer's integration surface rather than Kalcend's, and held Partial for want of a public API for Kalcend itself. That reasoning was right; the fact was not. THE PRIVACY POLICY ENUMERATES THE SERVICE'S COMPONENTS AND A DEVELOPER API IS ONE OF THEM: a SaaS platform for building and deploying AI agents with knowledge bases, third-party integrations, browser automation, automations, AND A DEVELOPER API. This is legal-tier evidence, which section 7 places at the top of the hierarchy precisely because counsel describes what the software actually does rather than what marketing wishes it did. A privacy policy has no incentive to invent an API. SIGNED WEBHOOKS ARE THE SECOND HALF and they run the other way. The Builder writes real integrations WITH SIGNED WEBHOOKS, and the privacy policy logs webhook activity as a distinct category. Signing is a detail a vendor implements when external systems are genuinely calling in, not a phrase invented for a features page. CONFIDENCE IS MEDIUM AND THE GAP IS NAMED PRECISELY: I did not reach an API reference, endpoint list, authentication documentation or SDK, so the existence of the developer API rests on the service description rather than on its documentation. No MCP server was found, which under the ruling does not withhold the grade. If a lane-close pass finds no reachable API documentation, this should return to Partial with that recorded. For a grant-stage vendor selling explicitly to non-technical buyers, shipping a developer API at all is worth noting; it is not where this product's audience lives.

Stands at F, and it is now among the better-evidenced Ext cells in the lane because a real API documentation site exists that the July build did not cite. THE DOCUMENTATION IS STRUCTURED AND SPECIFIC, not a contact-sales page. pickaxe.co/v1/documentation covers an Introduction, a COMPLETIONS API for deployment inference requests, CLI and Coding Agents, Prompt Injectors, Environment Variables, Webhooks and Authenticated Embeds, plus RESOURCE-SPECIFIC CRUD AND LISTING ENDPOINTS with request and response examples. CRUD across resources means the platform is administrable from outside, not merely queryable, which is the stronger form of what Mike's 30 August bar asks. THE CLI IS THE DISTINCTIVE PART AND IT IS AIMED SOMEWHERE UNUSUAL. It is described as BUILT FOR AI CODING AGENTS LIKE CLAUDE CODE, CURSOR, AND WINDSURF, letting a coding agent create agents, sync knowledge bases and deploy WITHOUT OPENING THE WEB APP. A no-code platform shipping a machine-first control surface is a deliberate second audience, and it is the Ext direction precisely: the customer's assistant operates Pickaxe. WEBHOOKS RUN OUTBOUND to Zapier, Make and other endpoints, and Authenticated Embeds cover programmatic session handoff, so the surface spans inbound control, outbound events and embedded delivery. ONE FACT RECORDED AND NOT CREDITED: a community-built MCP server exists on GitHub exposing agents, knowledge bases, users and analytics to Claude. It is third-party rather than the vendor's, so it carries nothing, but its existence corroborates that the public API is complete enough for someone outside the company to wrap it. No first-party MCP server was found, which under the ruling does not withhold the grade.

Stands at P under the 31 August Eval bar, and it now holds this fact exclusively rather than sharing it with two other axes. WHAT IS DOCUMENTED IS A REAL BUILD-TIME LOOP, and the features page states it as part of the build sequence: the Builder learns the business, generates the knowledge base, writes the integrations, and TESTS EVERY BUILD IN A LIVE PLAYGROUND BEFORE IT SHIPS. EVERY BUILD is the phrase that matters. Testing is not an optional step a diligent customer takes; it is in the path, which is a stronger commitment than most vendors at this grade make. The sandboxed playground lets teams simulate inputs and completions and fine-tune behaviour before deploying, so the customer can exercise the agent against their own scenarios rather than trusting a generated artefact blind. WHY IT STAYS PARTIAL. The ruled bar requires a result the customer can READ AND COMPARE about the agent's behaviour on their own work. A live playground shows what the agent does; it does not score it, retain the scenarios as a test set, or let a customer establish whether version two behaves better than version one after a prompt or knowledge change. Simulate-and-observe is a quality gate with no readable result, which is the ruled definition of Partial and fits precisely. THIS IS THE HONEST CEILING FOR THE PRODUCT'S DESIGN. In a platform where the Builder regenerates an agent from a prompt, the natural regression question, did my change help, is exactly the one nothing here answers. A scored playground with retained scenarios would take this to Full and is a small step from what exists; it is the first thing to check at lane close.

Stands at P under the 31 August Eval bar, and the reasoning is worth recording because this record has more testing surface than most Partials. WHAT EXISTS IS A REAL PRE-DEPLOYMENT LOOP. Preview and test mode, IMPERSONATING SPECIFIC USERS to test access paths, one-click model switching for comparison, a cost estimator, and AI-POWERED PROMPT IMPROVEMENTS WITH REAL-TIME TESTING. User impersonation is the standout: for a platform whose agents sit behind access groups and paywalls, testing what a member sees versus an invite-only user is exactly the failure mode that would otherwise reach a paying customer. THE OPTIMISATION LIMB IS GENUINELY SERVED. Comparing models on cost, speed and quality with live rates in view is optimisation with a readable number attached, and it is more than most vendors at this grade offer. WHY IT STOPS SHORT OF FULL. The ruled bar asks for a readable comparable result about THE AGENT'S BEHAVIOUR on the customer's own work. Cost and latency are properties of the model, not judgements about whether the agent answered correctly. Nothing scores an output, retains a test set, records expected answers, or lets a creator establish whether a prompt or knowledge change improved behaviour rather than merely changing it. Real-time testing shows what happens; it does not tell you whether it was better. THE GAP IS SHARPER HERE THAN THE GRADE CONVEYS, and worth carrying. Creators sell these agents to paying end users and the knowledge base re-syncs daily, so behaviour drifts without anyone changing anything. A platform with automatic content refresh and no regression check is one where quality can degrade silently between deployments. The Activity tab's conversation summaries support after-the-fact review and are credited on Obs.

N>F, and this is the largest single correction of the session. The July basis said no browser or computer use was documented. It is documented twice, on two different first-party pages, and one of them is the privacy policy. THE CAPABILITY STATEMENT: the features page lists, among the agent's abilities, NAVIGATES, ACTS, AND EXTRACTS, HANDING OFF TO A HUMAN FOR SIGN-IN OR OTP. Navigate and act are operation, not retrieval, and the one-time-passcode detail is the tell that this drives real authenticated sites rather than fetching public pages. THE CONFIRMATION IS LEGAL-TIER AND THAT IS WHY THIS IS FULL RATHER THAN PARTIAL. The privacy policy describes the service as comprising knowledge bases, third-party integrations, BROWSER AUTOMATION, automations and a developer API, and separately defines a category of processed data as tasks agents perform INCLUDING WEB PAGES AND DATA ACCESSED VIA BROWSER AUTOMATION. Counsel enumerating browser automation as a distinct activity is the strongest form of evidence in the hierarchy, because it describes what the software actually does and carries legal consequence if wrong. WHY IT IS THE POSITIVE CASE RATHER THAN THE NEAR-MISS I HAVE REFUSED REPEATEDLY. Kalcend ships 43 one-click API integrations. Browser automation exists alongside them, for the systems those integrations do not reach, which is the axis definition exactly: operating software the vendor does not control because no programmatic interface exists. The scraping refusals on ai-library and snaplogic were retrieval; this is navigation and action. WORTH RECORDING FOR THE LANE: this is a grant-stage single-founder vendor holding a genuine Comp cell in a lane where I have refused this axis on almost every larger platform. It is a useful counterweight to the assumption that computer use tracks company size. One limit named: browser only, with no desktop or operating-system control documented, and no statement of scope or reliability.

Stands at N, and confidence rises to high because this pass found two near-misses and refused both, which is a stronger basis than the July one of having seen no claim. NEAR-MISS ONE, ALREADY REFUSED IN JULY AND CORRECTLY: a Google search Action gives real-time web retrieval. Fetching results is retrieval, not operating an interface, and the July basis said so. NEAR-MISS TWO IS NEW AND IS THE ONE THAT WOULD CATCH A GRADER SKIMMING. The user manual lists CODE EXECUTION, RUN CODE DIRECTLY FROM THE AGENT, alongside PDF generation. Code execution is the most frequently miscounted fact on this axis and it was the exact defect corrected across thirteen records in the Coding agent June cohort. Running code is not operating software that lacks a programmatic interface; it is using the most programmatic interface there is. EVERYTHING ELSE RUNS THROUGH DOCUMENTED PROGRAMMATIC PATHS: more than 500 Actions and MCP servers, webhooks, the Completions API, and a vector knowledge base. The platform's entire premise is that a non-technical creator connects tools that already expose APIs. THE ABSENCE IS STRUCTURAL RATHER THAN A GAP, and worth stating so the cell is not misread as a deficiency. Pickaxe agents are conversational products embedded in a page, a portal, Slack or WhatsApp, answering a person. Driving a third-party interface has no place in that shape, and no page reached makes any claim in that territory. Recorded for the lane: this is the third record this session where code execution or scraping presented as computer use and was refused, against kalcend, where genuine browser navigation with an OTP handoff earned Full. The line has been applied identically in both directions.