Spaces:
Running
Running
Commit History
Today
-
Smart engine: parallel setting groups and step overlap off by default (fast tokenizers are not thread-safe: 'Already borrowed')
unknown +1: Claude Opus 5.5Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>5f3564c
Yesterday
-
Pay as you go only: prepaid wallet with PhonePe top-ups; every run reserved before it starts, capped (Groq + deadline) and charged its actual price with a 300% margin after GST and fees; admin profit report
unknown +1: Claude Opus 5.5Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>89d4f87 -
Smart Humanize faster: setting groups of a pass rewritten at once (GPU), rescue rewrites requested at the start of each call, vocabulary step alongside Parrot; job time in the log
unknown +1: Claude Opus 5.5Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>82a89bf
3 days ago
-
Paraphrasing 10x faster: learned banned words blocked with suppress_tokens
unknown +1: Claude Opus 5.5The adaptive learner's banned-word list had grown to 5,528 token sequences on the Space. Passed as bad_words_ids it is checked in Python at every generated token (with a GPU sync each time): every Parrot call took 18-28 s, even for one sentence, and it got slower as the tool learned. Single-token bans now go through suppress_tokens (one vectorised step); banned words of several tokens are still rejected afterwards by banned_in(). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>23008bb -
Log the time of each paraphrase generate call
unknown +1: Claude Opus 5.5Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>7be72fa -
Finer stage timing in the Space log
unknown +1: Claude Opus 5.5Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>33aab58 -
Result message names corrections held back by the proofread; stage timing in the Space log
unknown +1: Claude Opus 5.5Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>c7d7f21
8 days ago
-
Automatic proofread, pay-per-use charges with a pay pop-up, yearly plans
unknown +1: Claude Opus 5.5Automatic proofread - Every humanized text / document (Smart Humanize, Humanizer) goes through the final check & proofread; the proofread file is delivered. AUTO_PROOFREAD=0 = only on request. - humanizer.PROOFREAD_FOLLOWS: the humanizer's own language-model polish is skipped when the proofread follows (one final check, not two). - A failed proofread still delivers the humanized result and says so. Pay per use (plans.py, app.py) - check_doc returns "payg" for a signed-in customer without credits instead of blocking (PAY_PER_USE=0 restores the block). - usage_price(): GPU seconds x COST_GPU_HOUR_USD, idle share of the Space (COST_IDLE_FACTOR), the job's language-model cost, then PRICE_MARGIN (200%), PRICE_GST (18%), PRICE_USD_INR, PRICE_MIN_INR. - charges table; the result is locked (no file, no preview, no email, not in the run history) until settle_charge(ref). Pop-up with the price; Pay opens PAY_URL or the WhatsApp contact; "I have paid" unlocks. - Admin: pay-per-us…ca97939 -
Cut language-model cost and show what each job costs
unknown +1: Claude Opus 5.5Measured on the paid tier: one pass of every task over 5,000 words is about 6-7 US cents; the natural-style rewrites are over half of it. - Models: default list is gpt-oss-120b, gpt-oss-20b. qwen3.8-27b (about six times the price) and the unavailable llama model are no longer fallbacks. - Routing: verify and grammar run on the half-price GROQ_LIGHT_MODEL (GROQ_LIGHT_KINDS); writing tasks stay on the main model. - verify(): verdicts cached and packed through the shared batch runner. - Final polish is sent only for paragraphs the tool changed (POLISH_CHANGED_ONLY). - A rate limit that clears within GROQ_SHORT_LIMIT seconds is waited out instead of switching writing tasks to the weaker model. - GROQ_JOB_BUDGET: ceiling per job in US$ (default 1.00); past it the job continues with the local checks. - Accounting: every request booked by task (tokens in, cached, out, cost; GROQ_PRICES); "[cost]" line per Smart Humanize and proofread job in the log. - README: measured costs a…2906069 -
Merge the Space's 8-9 October work (origin/main) into the local changes
unknown +1: Claude Opus 5.5The Space had moved on (adaptive learning, detector-guided rewriting, token-budget handling in groqai, shared integrity checks, proofread changes). Conflicts in groqai, humanizer and textparse were resolved by keeping both sides: - groqai: the Space's token-budget chat loop, with the per-model fallback on a rate limit added to it. - humanizer: the Space's PARA_MODEL fallback inside the lazy load_model(); the automatic deeper rewrite (BOOST) feeds the adaptive candidate count and the plain-wording source. - textparse: both sets of guard checks run; strict_check keeps its rejection reasons and gains the Space's integrity checks. - integrity_issue: "bracket type changed" no longer fires when an author-year citation is dropped (it has no "["), so the guard can re-attach the citation again. 53 unit tests pass; text and Word end-to-end runs complete on the merge. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>eaf69a7 -
Make Smart Humanize a one-button screen; choose the rewrite depth automatically
unknown +1: Claude Opus 5.5- app: the settings panel is replaced by a short "paste or upload, press one button" intro. All sliders and checkboxes remain, folded into a closed "Advanced options (not needed)" section below the work area; the Humanizer tab's settings are folded the same way. The checker report is one optional collapsed line. - report_flags: the kind of report (AI / similarity) is recognised from the report itself; the radio defaults to "Recognise automatically". - smart + humanizer: when a pass does not lower the score, the sentences still flagged get the deeper rewrite for the remaining passes (humanizer.BOOST, SMART_AUTO_DEEP=0 = off), so the customer no longer has to tick "100% AI-written". Sentences that read as human stay untouched. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>726e1ce -
Add meaning check, Groq model fallback and subject test run; close guard gaps
unknown +1: Claude Opus 5.5Found by running Smart Humanize on 16 subject samples (devtools/batch.py). - groqai: on a rate limit the request moves to the next model in GROQ_MODELS (limits are per model), then back when the limit has passed. - groqai.verify + humanizer: every rewritten sentence is read against the author's sentence; one that changed the meaning (moved p-value, reversed comparison, dropped method, "but" -> "so") gets the next version that passes, else the original (HUM_VERIFY, HUM_VERIFY_ALTS). - textparse.guard: integrity rules now apply to every version; damaged citations are repaired or the rewrite rejected; each citation exactly once; bracket spacing normalised before the checks. - basic_grammar no longer treats "(Mishra and Mishra, 2012)" as a doubled word. - Sentence splitting: initials and case names ("Justice K.S. Puttaswamy v. Union of India") no longer end a sentence; detector and humanizer share one rule. - pubrules: micro-prefixed and compound units (ug/m3, mmol/g, MP…689da70 -
Harden rewrite integrity, stop the final polish undoing humanization, add tests
unknown +1: Claude Opus 5.5Integrity (textparse, pubrules) - Guard rejects lost relation signs (p < 0.01 -> p 0.01), duplicated or invented citations and unbalanced brackets; closes sentences left open. - New rules: no added cause, no added "we"; mediate/moderate/correlate/ regress protected in any word form. - Fewer false rejections: "indicate" is a reporting verb, not a hedge; another form of an author's name is not invented. - Author's typography kept (no-break hyphens, curly quotes, minus sign); fragments keep their open ending and lower-case start. - strict_check() returns the rejection reason. Humanizer / Smart Humanize - Final polish repairs only sentences that fail the clarity checks; a clear humanized sentence is no longer reverted to the original, with a paragraph-level check on optional edits. - Rescue versions are written in four distinct manners and chosen next to their neighbours' current wording. - Word: bold/italic runs inside a rewritten sentence keep their format; no full stop …585f232
10 days ago
-
Upload 5 files
Verified signature6600bf6 -
Upload 8 files
Verified signature0fdaff6 -
Upload 3 files
Verified signaturec351aaa
13 days ago
-
Upload 2 files
Verified signatured1763e8 -
Upload 2 files
Verified signature26dd9cb -
Upload 9 files
Verified signature78dbb12 -
Upload 4 files
Verified signature626fb61
14 days ago
-
Upload 4 files
Verified signaturef32a1d1 -
Upload 3 files
Verified signaturefa46575 -
Upload humanizer.py
Verified signatureacb385d -
Upload 5 files
Verified signature593e4c1 -
Upload 4 files
Verified signatureac1b38f -
Upload 4 files
Verified signature3655e6d -
Upload 5 files
Verified signature91003c8
16 days ago
-
Upload 6 files
Verified signature12739bc
17 days ago
-
Upload 3 files
Verified signaturecba1650 -
Upload 3 files
Verified signature418ece4 -
Upload 3 files
Verified signature2e843bb -
Upload 3 files
Verified signature1de19f4 -
Upload 4 files
Verified signature8dabb01 -
Upload 4 files
Verified signature2a58020 -
Upload 5 files
Verified signature77786c6 -
Upload 6 files
Verified signature473e146 -
Upload 5 files
Verified signature1d894db
18 days ago
-
Upload 2 files
Verified signature705055f -
Upload 2 files
Verified signaturee2f0d79 -
Upload 4 files
Verified signaturec4f9605 -
Upload 3 files
Verified signature18647a2
19 days ago
-
Upload 2 files
Verified signaturebc78f19 -
Upload 3 files
Verified signature17c6ef2 -
Upload 3 files
Verified signaturea7652c9 -
Upload 4 files
Verified signatureef25ab8 -
Upload 2 files
Verified signature2bb56a7 -
Upload humanizer.py
Verified signature7f9b4c2 -
Upload 3 files
Verified signature284649d -
Upload 5 files
Verified signaturec356456 -
Upload 20 files
Verified signature89d8bab