- GET /api/marketplace/personas/:id — returns all fields of a published persona
- Expand _COPY_FIELDS to include all new schema fields (oceanTraits, scenarios, aiSynthesizedBio, etc.)
- New MarketplacePersonaView page (/marketplace/personas/:id) with full sidebar + tabs
- Preview button navigates to full page instead of opening a summary dialog
- Clicking card body also navigates to persona page
- scripts/enrich_old_personas.py — LLM enrichment for 12 pre-schema personas
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Port fixes from Semblance:
- PERSONA_ALLOWED_FIELDS: 9 → 62 fields — oceanTraits, frustrations,
motivations, scenarios, aiSynthesizedBio, household/lifestyle fields,
audience_brief, research_objective, etc. were silently discarded at
Persona.create(), causing all OCEAN traits to render as 50% on frontend
- GET /personas/library — shared library endpoint (all users, bulk user lookup)
- POST /personas/:id/clone — clone any persona to current user
- GET /personas/:id — remove 403 for non-owners (read is public to auth users)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- RoundTable3D: replace sphere avatars with head+body person figures,
add per-persona glow pads, expanding pulse rings for active speaker,
connection line to table center, ambient particles, pill nameplates
with occupation, rotating center emblem
- FocusGroupSession: fix Rules-of-Hooks (move useMemo before early
returns), fix personas not loading (data.personas vs data.participants)
- Dark theme: replace hardcoded Tailwind colors across 20+ components
with semantic CSS tokens (bg-card, text-primary, text-muted-foreground,
text-brand-success, text-destructive, etc.)
- Add three + @types/three dependencies for 3D rendering
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- PDF: meta_row redesigned to same-line two-column layout so values
don't overflow the right edge; explicit cursor reset before each
multi_cell body() call fixes key themes appearing to the right of
titles; page numbers now use len(pdf.pages) + pdf.h-12 positioning
- Credits: replace fragile 4h credit_transactions lookup with atomic
findOneAndUpdate stamp on the focus group doc itself (24h window),
with rollback on insufficient balance — eliminates double-charging
on crash/restart; CreditTransaction.record failure is now non-fatal
- Key themes: cap input at 80 messages + max_output_tokens=4096 to
fix truncated JSON (Unterminated string at char 1580)
- Decision engine: require ≥4 participant responses per question before
moving on; mandate debate/contrarian seeking after 2 agreements;
call all participants to each question before anyone speaks 3rd time
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Old URL was cached by Telegram/Slack/etc with Lovable branding.
New URL forces a fresh fetch of the correct Cohorta image.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- _apply_silent_participant_override: new Python-level override that forces
any participant who hasn't spoken yet into the next turn (fires before
contrarian check, language-aware Russian/English call-out)
- _execute_participant_respond: now sends call_out as a moderator message
before generating the participant response (was silently skipped before,
causing 0-action loops and incoherent conversation flow); uses .get()
instead of [] to avoid KeyError when LLM omits optional fields
- _apply_contrarian_override: language-aware call-out message (Russian if
last moderator message contains Cyrillic)
- conversation-decision-engine.md: explicit rules that call_out MUST name
the participant and topic_context MUST give a specific angle; silent
participants must be called before repeat speakers
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- focus_group_ai.py: add missing `await` on all FocusGroup.add_message() calls
(was storing coroutine as task result → JSON serialization 500 error)
- model_pricing.py: remove broken MODEL_ALIASES import from llm_service
(was ImportError on every LLM usage cost tracking call)
- focus-group-response.md: language detection now based on last AI Moderator
message in conversation, not abstract topic field (fixes mixed-language responses)
- conversation_context_service.py: analytics now shows per-participant message
counts, explicit ⚠️ SILENT PARTICIPANTS and DOMINANT SPEAKERS warnings so
decision engine correctly engages quiet members and asks probing questions
- focus_groups.py PDF report: add full transcript section (new page), fix bar
chart to cap width within page bounds, add page number footers
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Restart endpoint now $unsets moderator_position, completion_reason,
autonomous_ended_at, autonomous_started_at so the AI doesn't
immediately conclude on the next run (was reading stale "guide complete" position)
- Duplicate endpoint excludes same fields from the copy for the same reason
- ParticipantPanel: add isEditable mode with remove buttons and
searchable "Add Participant" dialog (only active when status=new)
- FocusGroupSession: wire allPersonas state + add/remove handlers,
pass isEditable={status === 'new'} to ParticipantPanel
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Preview button on each persona card opens a dialog with description,
personality, age/gender/location/category badges, and buy button
- Publish toggle in admin now sends price+category alongside published flag,
so price is persisted even if admin never blurred the price input
- Fix price default in admin row: 0 instead of 5 to reflect actual DB state
- Fix confirmation dialog: price ?? 0 to avoid blank price display
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- New /api/marketplace/personas GET (list published) + POST /<id>/purchase (deduct credits, copy persona)
- Admin routes: GET /api/admin/marketplace/personas (all originals) + PUT /<id> (price/category/publish)
- Duplicate protection: can't buy same persona twice (409), can't publish purchased copies
- MarketplaceTab.tsx — admin table with inline price/category edit and publish toggle
- PersonaMarketplace.tsx — user-facing grid with search, category filter, buy dialog
- /marketplace route added to App.tsx + nav link in AppLayout
- Translations: en/ru/uk for all marketplace strings
- marketplace field added to PERSONA_ALLOWED_FIELDS
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Store amount_usd in Stripe webhook transaction (was only storing credits)
- Analytics endpoint: sum amount_usd for real USD revenue instead of credits
- Persona count: query personas collection directly (not credit transactions
which counted transactions, not individual personas — batch creation
of 3 at once appeared as count=1)
- Frontend: Revenue card now shows USD not credits; Gross Margin uses
USD revenue so calculation is dimensionally correct
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Per-persona message history: each persona now sees their own 8 previous
responses, preventing repetition and enabling position evolution
- OCEAN archetype labels in decision engine context: instead of raw numbers,
the decision LLM now sees "agreeableness: 72/100 [HIGH] — consensus-seeker"
- P2P interaction context: when participants interact directly, each one now
knows who they are responding to and what that person last said
- Python-level contrarian override: when agreement ratio in recent messages
exceeds 6% and a contrarian persona (low agreeableness or high neuroticism)
hasn't spoken recently, Python overrides moderator/probe action to call them
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add techSavviness, hasPurchasingPower, hasChildren, description to
_format_persona_details() — previously ignored despite being in persona model
- Rewrite OCEAN block: each trait now includes its psychological definition,
full name, and [LOW/MODERATE/HIGH] label so LLM understands the scoring scale
(e.g. "Agreeableness (cooperation, empathy, conflict avoidance): 25/100 [LOW]")
- Add extraversion behavioral constraint to _generate_behavioral_instructions()
(was used for response length only, never for personality guidance)
- Add LOW conscientiousness constraint (spontaneous, gut-feel)
- Add BALANCED PERSONALITY baseline for personas with all-moderate OCEAN scores —
previously these got zero behavioral constraints, behaving generically
- Add optimist/enthusiast personality keyword mapping
- Add OCEAN archetype legend to conversation-decision-engine.md so decision
engine understands what high/low scores mean when selecting next speaker
- Add +15% Diversity Boost modifier: prefer speakers whose archetype differs
from the last 2 participants (reduces echo-chamber dynamics)
- Add OCEAN model explanation header to focus-group-response.md prompt
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Extend _format_persona_details() to include background, traits, values,
interests, communication_style, brandLoyalty, priceConsciousness, demographics
(race/ethnicity/culture) — previously these fields were stored but never
sent to the LLM, causing all personas to give generic homogeneous answers
- Add _generate_behavioral_instructions() that translates OCEAN scores into
explicit behavioral directives (low agreeableness → must push back,
high neuroticism → voice of doubt, skeptic keyword → find the overpromise)
- Update focus-group-response.md to render BEHAVIORAL CONSTRAINTS and
DEMOGRAPHIC CONTEXT sections so LLM uses cultural/financial context
- Add anti-groupthink rule to conversation-decision-engine: on convergence,
prefer a contrarian persona before asking moderator probe question
- Improve ai-moderator-system.md: active @mention pull-in of silent
participants, spontaneous follow-up probes on interesting responses,
engagement tracking
- Add LANGUAGE RULE to all three response prompts: always match the language
of the question (fixes mixed Russian/English responses)
- Add POST /focus-groups/<id>/restart endpoint: clears session collections,
resets status to 'new', preserves config
- Add POST /focus-groups/<id>/duplicate endpoint: copies config to new group
- Frontend: Duplicate + Restart buttons on focus group cards with confirmation
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
dict(session.metadata) fails with KeyError:0 — metadata is a StripeObject
in SDK v5, not a plain dict. Use getattr() for all metadata fields.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
stripe.Webhook.construct_event() now returns StripeObject instances that
do not support .get()/__getitem__ — must use attribute access.
event["type"] → event.type, session.get("x") → session.x, etc.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
useMyUsage (and all other queries keyed without user identity) were showing
cached data from the previous user after logout/re-login. queryClient.clear()
in clearAuthData() wipes all in-memory cache on sign-out.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- /api/usage/me now returns credits_balance, credits_spent (MTD debits), and
last 30 credit transactions; no dollar amounts or token counts exposed
- MyUsage page redesigned: balance + MTD spent cards, transaction history table,
Buy Credits button
- Admin: POST /api/billing/test-checkout — $1 Stripe Checkout for any credit
amount to validate the payment flow end-to-end
- stripe_service: cast unit_amount to int (Stripe rejects float)
- i18n updated in EN/RU/UK
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- _get_runtime_config(): reads active provider endpoint, api_key, main/mini
model from app_settings (60s cache), falls back to env vars
- get_azure_client() now async, accepts cfg dict
- All generate_* methods call _get_runtime_config() per invocation so DB
changes take effect without restart
- app_settings: _seed_from_env() backfills empty endpoint/api_key from env
vars on first load so the admin UI shows current values immediately
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Batch-resolves user_id strings to emails after the aggregation via a
single find() on the users collection.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- PDF via fpdf2 (DejaVu font from fonts-dejavu-core, Cyrillic support)
- Lang param (?lang=ru/en) passed from i18n.language in ThemesPanel
- Prompts updated: executive summary + new key-decisions prompt, both
write entirely in the requested language
- Analytics section: participation bar chart per participant
- Key themes with quotes rendered as side-bordered blocks
- Filename: ASCII slug for Content-Disposition + UTF-8 filename* fallback
- api.ts: handles PDF blob, decodes UTF-8 filename*, 120s timeout
- AnalyticsPanel, ThemeHighlighter: light-mode colors → dark tokens
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
FocusGroupSession.tsx: replace bg-slate-50 with bg-background so the page
matches the rest of the dark UI; swap remaining text-slate-* for muted tokens.
focus_group_ai.py: before deducting 40 cr for an AI mode run, check if this
focus group already has a charge within the last 4 hours. Skip deduction if
found — prevents double-billing when the server restarts mid-session.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- AppLayout: add user info + 'My account' + 'Sign out' to mobile hamburger menu
- Header: add 'Sign out' button to authenticated mobile menu (landing page)
- dist/og-image.png: sync with public/ (Cohorta-branded, replaces Lovable image)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Previously Register.tsx wrote auth_token directly to localStorage without
calling setToken/setUser in AuthContext. This made isAuthenticated=true
(computed from localStorage) while user=null (React state), so ProtectedRoute's
user.email_verified check was never reached.
- AuthContext: add setAuth(token, userData) that syncs both localStorage and React state
- Register: use setAuth() instead of direct localStorage write
- Register: add !registered guard to auto-redirect effect to prevent bypassing check-inbox screen
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Without .dockerignore, COPY . . was overwriting the npm ci install
with the server's stale node_modules, causing Rollup to pick up the
wrong @radix-ui/react-slot version.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Fixes Docker build failure: react-progress brought react-primitive@2.1.4
which needs createSlot from react-slot >=1.2.0. npm overrides ensure
single 1.2.4 copy is used everywhere, no nested conflicts.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
