feat: wire persistence into the web layer (check → archive → dossier)

This is the piece that turns a Pre-Publish-Prüfung into an actual
archived, provable record instead of a one-off form response.

extract.Client.Extract now returns Result{Facts, RawJSON} instead of
just Facts — RawJSON is the model's exact, unmodified JSON, which is
what belongs in extraction.payload (the audit trail), not a re-encoded
view through our own Facts struct. extract.ParsePayload reconstructs
Facts from a stored payload later, reusing the same parsing/validation
path Extract uses (including the enum guard), so a previously-saved
extraction can be read back exactly as it would have been the first
time.

internal/dossier.BuildContent no longer requires AssetHash: most checks
right now are caption-only (no image/video upload wired yet), and
inventing a placeholder hash for a nonexistent asset would itself be an
integrity problem in an evidence tool. Content shows "kein Asset
hinterlegt" instead.

internal/web gains a narrow Store interface (mirroring the Extractor
pattern — only the methods these handlers use, not the full
*store.Store) so its test suite stays network/DB-free via an in-memory
fake:

- POST /pruefen persists submission + extraction + findings and marks
  the submission "checked". A needsClarification result persists the
  extraction (there's something worth keeping) but no findings and no
  status change, and the template omits the archive option entirely.
- POST /veroeffentlichen re-derives Facts from the stored payload, hashes
  the canonical submission+facts+findings metadata, gets an RFC-3161
  timestamp, generates the PDF dossier to disk, and persists the
  evidence_package row before marking the submission "published".
- GET /dossier/{id} serves the generated PDF.

Tested end-to-end offline: a fake Timestamper builds a real, structurally
valid self-signed RFC-3161 response so the full check→archive→download
flow runs against an in-memory store, verifying the downloaded bytes are
an actual PDF and the dossier file lands on disk — without hitting a
real database, TSA, or the Claude API.

cmd/deklarix/main.go now wires store.Store, evidence.NewHTTPTimestamper
(TSA_URL, default FreeTSA), and DOSSIER_DIR (default "dossiers") into
web.NewServer alongside the extractor and rule set.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

CLAUDE.md

@@ -126,7 +126,10 @@ Fünf Tabellen, mehr braucht der MVP nicht:
 - `submission` — ein eingereichter Beitrag, Status, Zeitpunkte
 - `asset` — Bild oder Datei, Pfad, SHA-256
 - `extraction` — das JSON aus Stufe 1, Modellversion, Prompt-Version
-- `finding` — Ergebnis pro Regel: Regel-ID, Regel-Version, Schwere, Text
+- `finding` — Ergebnis pro Regel: Regel-ID, Regel-Version, Schwere,
+  Titel, Korrektur, Fundstellen (zum Zeitpunkt des Findings fixiert,
+  kein Verweis auf die aktuelle Regel-YAML — die kann sich inzwischen
+  geändert haben)
 - `evidence_package` — Dossier, Hash, Zeitstempel-Token, erzeugt am
 - `participant` — Beteiligter an einer Submission mit Rolle
   (`creator`, `agentur`, `marke`, `kanzlei`) und Beitrag zur
@@ -330,7 +333,8 @@ git push origin main
 sudo systemctl start deklarix
 sudo systemctl status deklarix
 
-# Config: /etc/deklarix/deklarix.env (DATABASE_URL, PORT, ANTHROPIC_API_KEY, RULES_DIR)
+# Config: /etc/deklarix/deklarix.env (DATABASE_URL, PORT, ANTHROPIC_API_KEY,
+# RULES_DIR, DOSSIER_DIR, TSA_URL)
 
 # Logs prüfen
 journalctl -u deklarix -f

cmd/deklarix/main.go

@@ -6,6 +6,7 @@ import (
 	"net/http"
 	"os"
 
+	"github.com/netcell-it/deklarix/internal/evidence"
 	"github.com/netcell-it/deklarix/internal/extract"
 	"github.com/netcell-it/deklarix/internal/rules"
 	"github.com/netcell-it/deklarix/internal/store"
@@ -47,7 +48,14 @@ func main() {
 	}
 	log.Printf("%d Regeln aus %s geladen", len(ruleSet), rulesDir)
 
-	server, err := web.NewServer(extractor, ruleSet)
+	timestamper := evidence.NewHTTPTimestamper(os.Getenv("TSA_URL"))
+
+	dossierDir := os.Getenv("DOSSIER_DIR")
+	if dossierDir == "" {
+		dossierDir = "dossiers"
+	}
+
+	server, err := web.NewServer(extractor, ruleSet, db, timestamper, dossierDir)
 	if err != nil {
 		log.Fatalf("web server: %v", err)
 	}

internal/dossier/content.go

@@ -101,9 +101,11 @@ func BuildContent(data Data) (Content, error) {
 	if len(data.TimestampToken) == 0 {
 		return Content{}, fmt.Errorf("dossier: timestamp token fehlt")
 	}
-	if len(data.AssetHash) == 0 {
+	// AssetHash ist optional: nicht jede Prüfung hat ein hochgeladenes
-		return Content{}, fmt.Errorf("dossier: asset hash fehlt")
+	// Bild/Video (aktuell reine Caption-Prüfungen). Ein erfundener
-	}
+	// Platzhalter-Hash für ein nicht existierendes Asset wäre selbst ein
+	// Integritätsproblem — also zeigt das Dossier stattdessen "kein
+	// Asset hinterlegt" an, siehe Render.
 	if len(data.MetadataHash) == 0 {
 		return Content{}, fmt.Errorf("dossier: metadata hash fehlt")
 	}

internal/dossier/content_test.go

@@ -115,20 +115,28 @@ func TestBuildContentRejectsMissingTimestampToken(t *testing.T) {
 	}
 }
 
-func TestBuildContentRejectsMissingHashes(t *testing.T) {
+func TestBuildContentRejectsMissingMetadataHash(t *testing.T) {
 	data := validData(t)
-	data.AssetHash = nil
-	if _, err := dossier.BuildContent(data); err == nil {
-		t.Fatal("expected error for missing asset hash, got nil")
-	}
-
-	data = validData(t)
 	data.MetadataHash = nil
 	if _, err := dossier.BuildContent(data); err == nil {
 		t.Fatal("expected error for missing metadata hash, got nil")
 	}
 }
 
+func TestBuildContentAllowsMissingAssetHash(t *testing.T) {
+	// Reine Caption-Pruefungen ohne hochgeladenes Bild/Video haben kein
+	// Asset zum Hashen — das ist kein Fehlerfall.
+	data := validData(t)
+	data.AssetHash = nil
+	content, err := dossier.BuildContent(data)
+	if err != nil {
+		t.Fatalf("BuildContent: %v", err)
+	}
+	if content.AssetHashHex != "" {
+		t.Fatalf("AssetHashHex = %q, want empty when no asset was provided", content.AssetHashHex)
+	}
+}
+
 func TestBuildContentSuccess(t *testing.T) {
 	data := validData(t)
 

internal/dossier/render.go

@@ -89,7 +89,11 @@ func Render(w io.Writer, c Content) error {
 	pdf.Ln(2)
 
 	heading("Beweiskette")
+	if c.AssetHashHex != "" {
 		line("Asset-Hash (SHA-256): %s", c.AssetHashHex)
+	} else {
+		line("Asset-Hash: kein Asset hinterlegt (reine Caption-Prüfung)")
+	}
 	line("Metadaten-Hash (SHA-256): %s", c.MetadataHashHex)
 	line("RFC-3161-Zeitstempel: %s", c.TimestampedAt.Format("02.01.2006 15:04:05 MST"))
 	pdf.Ln(4)

internal/extract/client.go

@@ -46,6 +46,16 @@ type Input struct {
 	ImageData      []byte
 }
 
+// Result ist die Ausgabe von Extract: die für das Regelwerk
+// aufbereiteten Facts, plus RawJSON — das exakte, unveränderte JSON, das
+// das Modell zurückgegeben hat. RawJSON gehört unverändert in
+// extraction.payload (siehe Datenmodell); Facts ist eine abgeleitete
+// Sicht darauf und nicht der Beweis-Eintrag selbst.
+type Result struct {
+	Facts   rules.Facts
+	RawJSON []byte
+}
+
 // Client ruft die Claude API zur Fakten-Extraktion auf.
 type Client struct {
 	apiKey     string
@@ -100,15 +110,15 @@ func (c *Client) ModelVersion() string { return c.model }
 // ungültiger Enum-Wert) wird ein Fehler zurückgegeben statt stumm ein
 // Zero-Value-Facts zu liefern — ein falsches "keine Gegenleistung" wäre
 // hier schlimmer als ein sichtbarer Fehler.
-func (c *Client) Extract(ctx context.Context, in Input) (rules.Facts, error) {
+func (c *Client) Extract(ctx context.Context, in Input) (Result, error) {
 	if in.Caption == "" && len(in.ImageData) == 0 {
-		return rules.Facts{}, fmt.Errorf("extract: caption und bild sind beide leer")
+		return Result{}, fmt.Errorf("extract: caption und bild sind beide leer")
 	}
 
 	var content []contentBlock
 	if len(in.ImageData) > 0 {
 		if in.ImageMediaType == "" {
-			return rules.Facts{}, fmt.Errorf("extract: ImageMediaType fehlt für vorhandenes Bild")
+			return Result{}, fmt.Errorf("extract: ImageMediaType fehlt für vorhandenes Bild")
 		}
 		content = append(content, contentBlock{
 			Type: "image",
@@ -132,12 +142,12 @@ func (c *Client) Extract(ctx context.Context, in Input) (rules.Facts, error) {
 
 	payload, err := json.Marshal(reqBody)
 	if err != nil {
-		return rules.Facts{}, fmt.Errorf("extract: request marshal: %w", err)
+		return Result{}, fmt.Errorf("extract: request marshal: %w", err)
 	}
 
 	req, err := http.NewRequestWithContext(ctx, http.MethodPost, c.baseURL+"/v1/messages", bytes.NewReader(payload))
 	if err != nil {
-		return rules.Facts{}, fmt.Errorf("extract: request bauen: %w", err)
+		return Result{}, fmt.Errorf("extract: request bauen: %w", err)
 	}
 	req.Header.Set("content-type", "application/json")
 	req.Header.Set("x-api-key", c.apiKey)
@@ -145,20 +155,20 @@ func (c *Client) Extract(ctx context.Context, in Input) (rules.Facts, error) {
 
 	resp, err := c.httpClient.Do(req)
 	if err != nil {
-		return rules.Facts{}, fmt.Errorf("extract: request fehlgeschlagen: %w", err)
+		return Result{}, fmt.Errorf("extract: request fehlgeschlagen: %w", err)
 	}
 	defer resp.Body.Close()
 
 	var msg messageResponse
 	if err := json.NewDecoder(resp.Body).Decode(&msg); err != nil {
-		return rules.Facts{}, fmt.Errorf("extract: response decode: %w", err)
+		return Result{}, fmt.Errorf("extract: response decode: %w", err)
 	}
 
 	if resp.StatusCode != http.StatusOK {
 		if msg.Error != nil {
-			return rules.Facts{}, fmt.Errorf("extract: API-Fehler (%s): %s", msg.Error.Type, msg.Error.Message)
+			return Result{}, fmt.Errorf("extract: API-Fehler (%s): %s", msg.Error.Type, msg.Error.Message)
 		}
-		return rules.Facts{}, fmt.Errorf("extract: API-Status %d", resp.StatusCode)
+		return Result{}, fmt.Errorf("extract: API-Status %d", resp.StatusCode)
 	}
 
 	var toolUse *responseBlock
@@ -169,22 +179,43 @@ func (c *Client) Extract(ctx context.Context, in Input) (rules.Facts, error) {
 		}
 	}
 	if toolUse == nil {
-		return rules.Facts{}, fmt.Errorf("extract: keine tool_use-Antwort für %q enthalten", extractionTool.Name)
+		return Result{}, fmt.Errorf("extract: keine tool_use-Antwort für %q enthalten", extractionTool.Name)
 	}
 
+	facts, err := parseArgsToFacts(toolUse.Input, in.Platform, in.Jurisdiction)
+	if err != nil {
+		return Result{}, fmt.Errorf("extract: %w", err)
+	}
+
+	return Result{Facts: facts, RawJSON: toolUse.Input}, nil
+}
+
+// ParsePayload rekonstruiert Facts aus einem zuvor gespeicherten
+// RawJSON-Payload (z. B. aus extraction.payload). platform/jurisdiction
+// müssen erneut mitgegeben werden — sie sind, wie bei Extract, nie Teil
+// des Modell-Payloads.
+func ParsePayload(payload []byte, platform, jurisdiction string) (rules.Facts, error) {
+	facts, err := parseArgsToFacts(payload, platform, jurisdiction)
+	if err != nil {
+		return rules.Facts{}, fmt.Errorf("extract: %w", err)
+	}
+	return facts, nil
+}
+
+func parseArgsToFacts(raw []byte, platform, jurisdiction string) (rules.Facts, error) {
 	var args extractionArgs
-	if err := json.Unmarshal(toolUse.Input, &args); err != nil {
+	if err := json.Unmarshal(raw, &args); err != nil {
-		return rules.Facts{}, fmt.Errorf("extract: tool-input parse: %w", err)
+		return rules.Facts{}, fmt.Errorf("tool-input parse: %w", err)
 	}
 
 	consideration, err := parseConsideration(args.Consideration)
 	if err != nil {
-		return rules.Facts{}, fmt.Errorf("extract: %w", err)
+		return rules.Facts{}, err
 	}
 
 	return rules.Facts{
-		Platform:            in.Platform,
+		Platform:            platform,
-		Jurisdiction:        in.Jurisdiction,
+		Jurisdiction:        jurisdiction,
 		Consideration:       consideration,
 		DisclosurePresent:   args.DisclosurePresent,
 		DisclosureWording:   args.DisclosureWording,

internal/extract/client_test.go

@@ -103,8 +103,15 @@ func TestExtractSuccess(t *testing.T) {
 		DisclosureWording:   "Werbung",
 		DisclosureBeforeCut: false,
 	}
-	if got != want {
+	if got.Facts != want {
-		t.Fatalf("Extract() = %+v, want %+v", got, want)
+		t.Fatalf("Extract().Facts = %+v, want %+v", got.Facts, want)
+	}
+	var rawArgs map[string]any
+	if err := json.Unmarshal(got.RawJSON, &rawArgs); err != nil {
+		t.Fatalf("RawJSON does not parse as JSON: %v", err)
+	}
+	if rawArgs["gegenleistung"] != "bezahlt" {
+		t.Fatalf("RawJSON = %s, expected it to contain the model's raw field names", got.RawJSON)
 	}
 
 	toolChoice, _ := gotBody["tool_choice"].(map[string]any)
@@ -131,8 +138,8 @@ func TestExtractPassesThroughUnclear(t *testing.T) {
 	if err != nil {
 		t.Fatalf("Extract: %v", err)
 	}
-	if got.Consideration != rules.ConsiderationUnclear {
+	if got.Facts.Consideration != rules.ConsiderationUnclear {
-		t.Fatalf("Consideration = %q, want unklar", got.Consideration)
+		t.Fatalf("Consideration = %q, want unklar", got.Facts.Consideration)
 	}
 }
 
@@ -174,6 +181,34 @@ func TestExtractRejectsMalformedToolInput(t *testing.T) {
 	}
 }
 
+func TestParsePayloadRoundTrip(t *testing.T) {
+	payload := []byte(`{"gegenleistung":"sachbezug","kennzeichnung_vorhanden":true,"kennzeichnung_wortlaut":"Anzeige","kennzeichnung_vor_kuerzung":true}`)
+
+	facts, err := extract.ParsePayload(payload, "tiktok", "DE")
+	if err != nil {
+		t.Fatalf("ParsePayload: %v", err)
+	}
+
+	want := rules.Facts{
+		Platform:            "tiktok",
+		Jurisdiction:        "DE",
+		Consideration:       rules.ConsiderationInKind,
+		DisclosurePresent:   true,
+		DisclosureWording:   "Anzeige",
+		DisclosureBeforeCut: true,
+	}
+	if facts != want {
+		t.Fatalf("ParsePayload() = %+v, want %+v", facts, want)
+	}
+}
+
+func TestParsePayloadRejectsInvalidConsideration(t *testing.T) {
+	payload := []byte(`{"gegenleistung":"vielleicht","kennzeichnung_vorhanden":false,"kennzeichnung_wortlaut":"","kennzeichnung_vor_kuerzung":false}`)
+	if _, err := extract.ParsePayload(payload, "tiktok", "DE"); err == nil {
+		t.Fatal("expected error for out-of-enum stored payload, got nil")
+	}
+}
+
 func TestExtractPropagatesAPIError(t *testing.T) {
 	c := newTestClient(t, func(w http.ResponseWriter, r *http.Request) {
 		w.WriteHeader(http.StatusTooManyRequests)

internal/web/handlers.go

@@ -1,9 +1,15 @@
 package web
 
 import (
+	"encoding/hex"
 	"fmt"
 	"net/http"
+	"os"
+	"path/filepath"
+	"time"
 
+	"github.com/netcell-it/deklarix/internal/dossier"
+	"github.com/netcell-it/deklarix/internal/evidence"
 	"github.com/netcell-it/deklarix/internal/extract"
 	"github.com/netcell-it/deklarix/internal/rules"
 )
@@ -33,13 +39,18 @@ type findingView struct {
 }
 
 type resultData struct {
+	SubmissionID       string
 	NeedsClarification bool
+	CanArchive         bool
 	Findings           []findingView
 }
 
-// handleCheck führt die Pre-Publish-Prüfung aus: Extraktion (Stufe 1)
+// handleCheck führt die Pre-Publish-Prüfung aus: Extraktion (Stufe 1),
-// gefolgt von der Regelauswertung (Stufe 2). Speichert noch nichts —
+// Regelauswertung (Stufe 2), und speichert Submission + Extraction +
-// das ist der nächste Schritt (Verdrahtung über internal/store).
+// Findings. Die eigentliche Archivierung (Hash, Zeitstempel, Dossier)
+// passiert erst in handleArchive, wenn der Beitrag tatsächlich
+// veröffentlicht wurde — das sind unterschiedliche Zeitpunkte im
+// Lebenszyklus (siehe CLAUDE.md, Funktion 1 vs. 2).
 func (s *Server) handleCheck(w http.ResponseWriter, r *http.Request) {
 	if err := r.ParseForm(); err != nil {
 		http.Error(w, "ungültiges Formular", http.StatusBadRequest)
@@ -47,13 +58,16 @@ func (s *Server) handleCheck(w http.ResponseWriter, r *http.Request) {
 	}
 
 	platform := r.FormValue("platform")
+	postType := r.FormValue("post_type")
 	caption := r.FormValue("caption")
-	if platform == "" || caption == "" {
+	if platform == "" || postType == "" || caption == "" {
-		http.Error(w, "Plattform und Caption sind Pflichtfelder", http.StatusBadRequest)
+		http.Error(w, "Plattform, Beitragstyp und Caption sind Pflichtfelder", http.StatusBadRequest)
 		return
 	}
 
-	facts, err := s.extractor.Extract(r.Context(), extract.Input{
+	ctx := r.Context()
+
+	result, err := s.extractor.Extract(ctx, extract.Input{
 		Platform:     platform,
 		Jurisdiction: "DE",
 		Caption:      caption,
@@ -63,21 +77,180 @@ func (s *Server) handleCheck(w http.ResponseWriter, r *http.Request) {
 		return
 	}
 
-	findings, needsClarification := rules.Evaluate(s.ruleSet, facts)
+	sub, err := s.store.CreateSubmission(ctx, platform, postType, caption)
+	if err != nil {
+		http.Error(w, "Beitrag konnte nicht gespeichert werden: "+err.Error(), http.StatusInternalServerError)
+		return
+	}
 
-	data := resultData{NeedsClarification: needsClarification}
+	ext, err := s.store.CreateExtraction(ctx, sub.ID, result.RawJSON, s.extractor.ModelVersion(), extract.PromptVersion)
+	if err != nil {
+		http.Error(w, "Extraktion konnte nicht gespeichert werden: "+err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	findings, needsClarification := rules.Evaluate(s.ruleSet, result.Facts)
+
+	data := resultData{SubmissionID: sub.ID, NeedsClarification: needsClarification}
+	if !needsClarification {
+		extractionID := ext.ID
 		for _, f := range findings {
+			if _, err := s.store.CreateFinding(ctx, sub.ID, &extractionID, f.RuleID, f.RuleVersion, string(f.Severity), f.Title, f.Fix, f.Sources); err != nil {
+				http.Error(w, "Finding konnte nicht gespeichert werden: "+err.Error(), http.StatusInternalServerError)
+				return
+			}
 			data.Findings = append(data.Findings, findingView{
-			RuleID:   f.RuleID,
+				RuleID: f.RuleID, Version: f.RuleVersion, Severity: string(f.Severity),
-			Version:  f.RuleVersion,
+				Title: f.Title, Fix: f.Fix, Sources: f.Sources,
-			Severity: string(f.Severity),
-			Title:    f.Title,
-			Fix:      f.Fix,
-			Sources:  f.Sources,
 			})
 		}
+		if err := s.store.SetSubmissionStatus(ctx, sub.ID, "checked"); err != nil {
+			http.Error(w, "Status konnte nicht aktualisiert werden: "+err.Error(), http.StatusInternalServerError)
+			return
+		}
+		data.CanArchive = true
+	}
 
 	if err := s.templates.ExecuteTemplate(w, "result", data); err != nil {
 		http.Error(w, "Ergebnis konnte nicht gerendert werden", http.StatusInternalServerError)
 	}
 }
+
+type archivedData struct {
+	SubmissionID  string
+	DossierURL    string
+	TimestampedAt string
+}
+
+// handleArchive markiert einen geprüften Beitrag als veröffentlicht und
+// archiviert ihn: Hash der Metadaten, RFC-3161-Zeitstempel, PDF-Dossier.
+// Setzt eine vorherige Prüfung voraus (Submission + Extraction +
+// Findings müssen bereits existieren).
+func (s *Server) handleArchive(w http.ResponseWriter, r *http.Request) {
+	if err := r.ParseForm(); err != nil {
+		http.Error(w, "ungültiges Formular", http.StatusBadRequest)
+		return
+	}
+	submissionID := r.FormValue("submission_id")
+	if submissionID == "" {
+		http.Error(w, "submission_id ist Pflicht", http.StatusBadRequest)
+		return
+	}
+
+	ctx := r.Context()
+
+	sub, err := s.store.GetSubmission(ctx, submissionID)
+	if err != nil {
+		http.Error(w, "Beitrag nicht gefunden: "+err.Error(), http.StatusNotFound)
+		return
+	}
+
+	ext, err := s.store.GetLatestExtraction(ctx, submissionID)
+	if err != nil {
+		http.Error(w, "Extraktion nicht gefunden: "+err.Error(), http.StatusNotFound)
+		return
+	}
+	facts, err := extract.ParsePayload(ext.Payload, sub.Platform, "DE")
+	if err != nil {
+		http.Error(w, "gespeicherte Extraktion konnte nicht gelesen werden: "+err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	storeFindings, err := s.store.ListCurrentFindings(ctx, submissionID)
+	if err != nil {
+		http.Error(w, "Findings konnten nicht gelesen werden: "+err.Error(), http.StatusInternalServerError)
+		return
+	}
+	dossierFindings := make([]rules.Finding, len(storeFindings))
+	for i, f := range storeFindings {
+		dossierFindings[i] = rules.Finding{
+			RuleID: f.RuleID, RuleVersion: f.RuleVersion, Severity: rules.Severity(f.Severity),
+			Title: f.Title, Fix: f.Fix, Sources: f.Sources,
+		}
+	}
+
+	metadataHash, err := evidence.HashMetadata(struct {
+		SubmissionID string
+		Platform     string
+		PostType     string
+		Caption      string
+		Facts        rules.Facts
+		Findings     []rules.Finding
+	}{sub.ID, sub.Platform, sub.PostType, sub.Caption, facts, dossierFindings})
+	if err != nil {
+		http.Error(w, "Metadaten-Hash fehlgeschlagen: "+err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	timestampToken, err := s.timestamper.Timestamp(ctx, metadataHash)
+	if err != nil {
+		http.Error(w, "Zeitstempel fehlgeschlagen: "+err.Error(), http.StatusBadGateway)
+		return
+	}
+
+	if err := os.MkdirAll(s.dossierDir, 0o750); err != nil {
+		http.Error(w, "Dossier-Verzeichnis konnte nicht angelegt werden: "+err.Error(), http.StatusInternalServerError)
+		return
+	}
+	dossierPath := filepath.Join(s.dossierDir, submissionID+".pdf")
+	dossierFile, err := os.Create(dossierPath)
+	if err != nil {
+		http.Error(w, "Dossier-Datei konnte nicht angelegt werden: "+err.Error(), http.StatusInternalServerError)
+		return
+	}
+	defer dossierFile.Close()
+
+	genErr := dossier.Generate(dossierFile, dossier.Data{
+		Submission: dossier.Submission{
+			Platform: sub.Platform, PostType: sub.PostType, Caption: sub.Caption, CreatedAt: sub.CreatedAt,
+		},
+		Facts:          facts,
+		Findings:       dossierFindings,
+		MetadataHash:   metadataHash,
+		TimestampToken: timestampToken,
+		GeneratedAt:    time.Now(),
+	})
+	if genErr != nil {
+		http.Error(w, "Dossier konnte nicht erzeugt werden: "+genErr.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	if _, err := s.store.CreateEvidencePackage(ctx, submissionID, dossierPath, hex.EncodeToString(metadataHash), timestampToken); err != nil {
+		http.Error(w, "Evidence-Package konnte nicht gespeichert werden: "+err.Error(), http.StatusInternalServerError)
+		return
+	}
+	if err := s.store.SetSubmissionStatus(ctx, submissionID, "published"); err != nil {
+		http.Error(w, "Status konnte nicht aktualisiert werden: "+err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	timestampedAt, err := evidence.TimestampTime(timestampToken)
+	if err != nil {
+		http.Error(w, "Zeitstempel konnte nicht gelesen werden: "+err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	data := archivedData{
+		SubmissionID:  submissionID,
+		DossierURL:    "/dossier/" + submissionID,
+		TimestampedAt: timestampedAt.Format("02.01.2006 15:04:05 MST"),
+	}
+	if err := s.templates.ExecuteTemplate(w, "archived", data); err != nil {
+		http.Error(w, "Ergebnis konnte nicht gerendert werden", http.StatusInternalServerError)
+	}
+}
+
+// handleDossierDownload liefert das erzeugte PDF-Dossier eines
+// archivierten Beitrags aus.
+func (s *Server) handleDossierDownload(w http.ResponseWriter, r *http.Request) {
+	id := r.PathValue("id")
+
+	pkg, err := s.store.GetLatestEvidencePackage(r.Context(), id)
+	if err != nil {
+		http.Error(w, "Kein Dossier für diesen Beitrag gefunden", http.StatusNotFound)
+		return
+	}
+
+	w.Header().Set("Content-Type", "application/pdf")
+	http.ServeFile(w, r, pkg.DossierPath)
+}

internal/web/server.go

@@ -14,8 +14,10 @@ import (
 	"html/template"
 	"net/http"
 
+	"github.com/netcell-it/deklarix/internal/evidence"
 	"github.com/netcell-it/deklarix/internal/extract"
 	"github.com/netcell-it/deklarix/internal/rules"
+	"github.com/netcell-it/deklarix/internal/store"
 )
 
 //go:embed templates/*.html
@@ -28,7 +30,24 @@ var staticFS embed.FS
 // *extract.Client erfüllt sie; Tests injizieren einen Fake statt echte
 // Claude-API-Aufrufe zu machen.
 type Extractor interface {
-	Extract(ctx context.Context, in extract.Input) (rules.Facts, error)
+	Extract(ctx context.Context, in extract.Input) (extract.Result, error)
+	ModelVersion() string
+}
+
+// Store ist die Schnittstelle, die der Server zur Persistenz braucht —
+// bewusst schmal (nur was diese Handler tatsächlich nutzen), nicht der
+// komplette *store.Store. *store.Store erfüllt sie; Tests injizieren
+// einen Fake statt eine echte Datenbank zu brauchen.
+type Store interface {
+	CreateSubmission(ctx context.Context, platform, postType, caption string) (store.Submission, error)
+	GetSubmission(ctx context.Context, id string) (store.Submission, error)
+	SetSubmissionStatus(ctx context.Context, id, status string) error
+	CreateExtraction(ctx context.Context, submissionID string, payload []byte, modelVersion, promptVersion string) (store.Extraction, error)
+	GetLatestExtraction(ctx context.Context, submissionID string) (store.Extraction, error)
+	CreateFinding(ctx context.Context, submissionID string, extractionID *string, ruleID string, ruleVersion int, severity, title, fix string, sources []string) (store.Finding, error)
+	ListCurrentFindings(ctx context.Context, submissionID string) ([]store.Finding, error)
+	CreateEvidencePackage(ctx context.Context, submissionID, dossierPath, sha256Hex string, timestampToken []byte) (store.EvidencePackage, error)
+	GetLatestEvidencePackage(ctx context.Context, submissionID string) (store.EvidencePackage, error)
 }
 
 // Server bündelt Routing und Abhängigkeiten der Web-Schicht.
@@ -36,23 +55,36 @@ type Server struct {
 	mux         *http.ServeMux
 	extractor   Extractor
 	ruleSet     []rules.Rule
+	store       Store
+	timestamper evidence.Timestamper
+	dossierDir  string
 	templates   *template.Template
 }
 
 // NewServer erstellt den Server. ruleSet kommt von rules.Load und wird
-// einmal beim Start geladen, nicht pro Request.
+// einmal beim Start geladen, nicht pro Request. dossierDir ist das
-func NewServer(extractor Extractor, ruleSet []rules.Rule) (*Server, error) {
+// Verzeichnis, in das erzeugte Nachweis-Dossiers geschrieben werden.
+func NewServer(extractor Extractor, ruleSet []rules.Rule, st Store, timestamper evidence.Timestamper, dossierDir string) (*Server, error) {
 	tmpl, err := template.ParseFS(templatesFS, "templates/*.html")
 	if err != nil {
 		return nil, fmt.Errorf("web: templates parsen: %w", err)
 	}
 
-	s := &Server{extractor: extractor, ruleSet: ruleSet, templates: tmpl}
+	s := &Server{
+		extractor:   extractor,
+		ruleSet:     ruleSet,
+		store:       st,
+		timestamper: timestamper,
+		dossierDir:  dossierDir,
+		templates:   tmpl,
+	}
 
 	mux := http.NewServeMux()
 	mux.HandleFunc("GET /health", s.handleHealth)
 	mux.HandleFunc("GET /{$}", s.handleIndex)
 	mux.HandleFunc("POST /pruefen", s.handleCheck)
+	mux.HandleFunc("POST /veroeffentlichen", s.handleArchive)
+	mux.HandleFunc("GET /dossier/{id}", s.handleDossierDownload)
 	mux.Handle("GET /static/", http.FileServerFS(staticFS))
 	s.mux = mux
 

internal/web/server_test.go

@@ -2,26 +2,217 @@ package web_test
 
 import (
 	"context"
+	"crypto"
+	"crypto/ecdsa"
+	"crypto/elliptic"
+	"crypto/rand"
+	"crypto/x509"
+	"crypto/x509/pkix"
+	"encoding/asn1"
+	"fmt"
+	"math/big"
 	"net/http"
 	"net/http/httptest"
 	"net/url"
 	"os"
 	"strings"
+	"sync"
 	"testing"
+	"time"
+
+	"github.com/digitorus/timestamp"
 
 	"github.com/netcell-it/deklarix/internal/extract"
 	"github.com/netcell-it/deklarix/internal/rules"
+	"github.com/netcell-it/deklarix/internal/store"
 	"github.com/netcell-it/deklarix/internal/web"
 )
 
+// ─── Fakes ────────────────────────────────────────────────────────────
+
 type fakeExtractor struct {
 	facts rules.Facts
+	raw   []byte
 	err   error
 }
 
-func (f fakeExtractor) Extract(ctx context.Context, in extract.Input) (rules.Facts, error) {
+func (f fakeExtractor) Extract(ctx context.Context, in extract.Input) (extract.Result, error) {
-	return f.facts, f.err
+	if f.err != nil {
+		return extract.Result{}, f.err
 	}
+	raw := f.raw
+	if raw == nil {
+		raw = []byte(`{}`)
+	}
+	return extract.Result{Facts: f.facts, RawJSON: raw}, nil
+}
+
+func (f fakeExtractor) ModelVersion() string { return "fake-model-v0" }
+
+// fakeStore ist eine In-Memory-Implementierung von web.Store, damit die
+// Handler-Tests keine echte Postgres-Instanz brauchen.
+type fakeStore struct {
+	mu           sync.Mutex
+	nextID       int
+	submissions  map[string]store.Submission
+	extractions  map[string]store.Extraction
+	findings     map[string][]store.Finding
+	evidencePkgs map[string]store.EvidencePackage
+}
+
+func newFakeStore() *fakeStore {
+	return &fakeStore{
+		submissions:  map[string]store.Submission{},
+		extractions:  map[string]store.Extraction{},
+		findings:     map[string][]store.Finding{},
+		evidencePkgs: map[string]store.EvidencePackage{},
+	}
+}
+
+func (f *fakeStore) newID() string {
+	f.nextID++
+	return fmt.Sprintf("id-%d", f.nextID)
+}
+
+func (f *fakeStore) CreateSubmission(ctx context.Context, platform, postType, caption string) (store.Submission, error) {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	sub := store.Submission{
+		ID: f.newID(), Platform: platform, PostType: postType, Caption: caption,
+		Status: "draft", CreatedAt: time.Now(), UpdatedAt: time.Now(),
+	}
+	f.submissions[sub.ID] = sub
+	return sub, nil
+}
+
+func (f *fakeStore) GetSubmission(ctx context.Context, id string) (store.Submission, error) {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	sub, ok := f.submissions[id]
+	if !ok {
+		return store.Submission{}, fmt.Errorf("fakeStore: submission %s nicht gefunden", id)
+	}
+	return sub, nil
+}
+
+func (f *fakeStore) SetSubmissionStatus(ctx context.Context, id, status string) error {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	sub, ok := f.submissions[id]
+	if !ok {
+		return fmt.Errorf("fakeStore: submission %s nicht gefunden", id)
+	}
+	sub.Status = status
+	f.submissions[id] = sub
+	return nil
+}
+
+func (f *fakeStore) CreateExtraction(ctx context.Context, submissionID string, payload []byte, modelVersion, promptVersion string) (store.Extraction, error) {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	ext := store.Extraction{
+		ID: f.newID(), SubmissionID: submissionID, Payload: payload,
+		ModelVersion: modelVersion, PromptVersion: promptVersion, CreatedAt: time.Now(),
+	}
+	f.extractions[submissionID] = ext
+	return ext, nil
+}
+
+func (f *fakeStore) GetLatestExtraction(ctx context.Context, submissionID string) (store.Extraction, error) {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	ext, ok := f.extractions[submissionID]
+	if !ok {
+		return store.Extraction{}, fmt.Errorf("fakeStore: keine Extraktion fuer %s", submissionID)
+	}
+	return ext, nil
+}
+
+func (f *fakeStore) CreateFinding(ctx context.Context, submissionID string, extractionID *string, ruleID string, ruleVersion int, severity, title, fix string, sources []string) (store.Finding, error) {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	finding := store.Finding{
+		ID: f.newID(), SubmissionID: submissionID, ExtractionID: extractionID,
+		RuleID: ruleID, RuleVersion: ruleVersion, Severity: severity,
+		Title: title, Fix: fix, Sources: sources, CreatedAt: time.Now(),
+	}
+	f.findings[submissionID] = append(f.findings[submissionID], finding)
+	return finding, nil
+}
+
+func (f *fakeStore) ListCurrentFindings(ctx context.Context, submissionID string) ([]store.Finding, error) {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	return f.findings[submissionID], nil
+}
+
+func (f *fakeStore) CreateEvidencePackage(ctx context.Context, submissionID, dossierPath, sha256Hex string, timestampToken []byte) (store.EvidencePackage, error) {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	pkg := store.EvidencePackage{
+		ID: f.newID(), SubmissionID: submissionID, DossierPath: dossierPath,
+		SHA256: sha256Hex, TimestampToken: timestampToken, CreatedAt: time.Now(),
+	}
+	f.evidencePkgs[submissionID] = pkg
+	return pkg, nil
+}
+
+func (f *fakeStore) GetLatestEvidencePackage(ctx context.Context, submissionID string) (store.EvidencePackage, error) {
+	f.mu.Lock()
+	defer f.mu.Unlock()
+	pkg, ok := f.evidencePkgs[submissionID]
+	if !ok {
+		return store.EvidencePackage{}, fmt.Errorf("fakeStore: kein evidence package fuer %s", submissionID)
+	}
+	return pkg, nil
+}
+
+// fakeTimestamper liefert einen offline erzeugten, strukturell gültigen
+// (selbstsignierten) RFC-3161-Token — genug, damit evidence.TimestampTime
+// ihn parsen kann, ohne eine echte TSA zu brauchen.
+type fakeTimestamper struct{}
+
+func (fakeTimestamper) Timestamp(ctx context.Context, hash []byte) ([]byte, error) {
+	key, err := ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
+	if err != nil {
+		return nil, err
+	}
+	now := time.Now()
+	certTemplate := &x509.Certificate{
+		SerialNumber: big.NewInt(1),
+		Subject:      pkix.Name{CommonName: "deklarix-test-tsa"},
+		NotBefore:    now.Add(-time.Hour),
+		NotAfter:     now.Add(time.Hour),
+		KeyUsage:     x509.KeyUsageDigitalSignature,
+		ExtKeyUsage:  []x509.ExtKeyUsage{x509.ExtKeyUsageTimeStamping},
+	}
+	certDER, err := x509.CreateCertificate(rand.Reader, certTemplate, certTemplate, &key.PublicKey, key)
+	if err != nil {
+		return nil, err
+	}
+	cert, err := x509.ParseCertificate(certDER)
+	if err != nil {
+		return nil, err
+	}
+	ts := timestamp.Timestamp{
+		HashAlgorithm: crypto.SHA256,
+		HashedMessage: hash,
+		Time:          now,
+		SerialNumber:  big.NewInt(1),
+		Policy:        asn1.ObjectIdentifier{1, 2, 3},
+	}
+	respDER, err := ts.CreateResponse(cert, key)
+	if err != nil {
+		return nil, err
+	}
+	parsed, err := timestamp.ParseResponse(respDER)
+	if err != nil {
+		return nil, err
+	}
+	return parsed.RawToken, nil
+}
+
+// ─── Test-Setup ───────────────────────────────────────────────────────
 
 func loadRealRules(t *testing.T) []rules.Rule {
 	t.Helper()
@@ -34,13 +225,15 @@ func loadRealRules(t *testing.T) []rules.Rule {
 
 func newTestServer(t *testing.T, ex web.Extractor) *web.Server {
 	t.Helper()
-	s, err := web.NewServer(ex, loadRealRules(t))
+	s, err := web.NewServer(ex, loadRealRules(t), newFakeStore(), fakeTimestamper{}, t.TempDir())
 	if err != nil {
 		t.Fatalf("NewServer: %v", err)
 	}
 	return s
 }
 
+// ─── Tests ────────────────────────────────────────────────────────────
+
 func TestHandleHealth(t *testing.T) {
 	s := newTestServer(t, fakeExtractor{})
 
@@ -67,7 +260,7 @@ func TestHandleIndexRendersForm(t *testing.T) {
 		t.Fatalf("status = %d, want 200", w.Code)
 	}
 	body := w.Body.String()
-	for _, want := range []string{`name="platform"`, `name="caption"`, `hx-post="/pruefen"`} {
+	for _, want := range []string{`name="platform"`, `name="post_type"`, `name="caption"`, `hx-post="/pruefen"`} {
 		if !strings.Contains(body, want) {
 			t.Errorf("index body missing %q\nbody: %s", want, body)
 		}
@@ -89,64 +282,87 @@ func TestHandleStaticServesHTMX(t *testing.T) {
 	}
 }
 
-func postCheck(t *testing.T, s *web.Server, form url.Values) *httptest.ResponseRecorder {
+func postForm(t *testing.T, s *web.Server, path string, form url.Values) *httptest.ResponseRecorder {
 	t.Helper()
-	req := httptest.NewRequest(http.MethodPost, "/pruefen", strings.NewReader(form.Encode()))
+	req := httptest.NewRequest(http.MethodPost, path, strings.NewReader(form.Encode()))
 	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
 	w := httptest.NewRecorder()
 	s.ServeHTTP(w, req)
 	return w
 }
 
+func checkForm(extra ...string) url.Values {
+	v := url.Values{"platform": {"instagram"}, "post_type": {"reel"}, "caption": {"..."}}
+	for i := 0; i+1 < len(extra); i += 2 {
+		v.Set(extra[i], extra[i+1])
+	}
+	return v
+}
+
 func TestHandleCheckRejectsMissingFields(t *testing.T) {
 	s := newTestServer(t, fakeExtractor{})
 
-	w := postCheck(t, s, url.Values{"platform": {"instagram"}})
+	w := postForm(t, s, "/pruefen", url.Values{"platform": {"instagram"}})
 	if w.Code != http.StatusBadRequest {
-		t.Fatalf("status = %d, want 400 for missing caption", w.Code)
+		t.Fatalf("status = %d, want 400 for missing fields", w.Code)
 	}
 }
 
 func TestHandleCheckPropagatesExtractionError(t *testing.T) {
-	s := newTestServer(t, fakeExtractor{err: errTest})
+	s := newTestServer(t, fakeExtractor{err: fmt.Errorf("simulierter Extraktionsfehler")})
 
-	w := postCheck(t, s, url.Values{"platform": {"instagram"}, "caption": {"x"}})
+	w := postForm(t, s, "/pruefen", checkForm())
 	if w.Code != http.StatusBadGateway {
 		t.Fatalf("status = %d, want 502 when extraction fails", w.Code)
 	}
 }
 
-func TestHandleCheckRendersFinding(t *testing.T) {
+func TestHandleCheckPersistsSubmissionExtractionAndFindings(t *testing.T) {
-	s := newTestServer(t, fakeExtractor{facts: rules.Facts{
+	fs := newFakeStore()
-		Platform:            "instagram",
+	s, err := web.NewServer(fakeExtractor{facts: rules.Facts{
-		Jurisdiction:        "DE",
+		Platform: "instagram", Jurisdiction: "DE", Consideration: rules.ConsiderationPaid,
-		Consideration:       rules.ConsiderationPaid,
+		DisclosurePresent: true, DisclosureWording: "Werbung", DisclosureBeforeCut: false,
-		DisclosurePresent:   true,
+	}, raw: []byte(`{"gegenleistung":"bezahlt"}`)}, loadRealRules(t), fs, fakeTimestamper{}, t.TempDir())
-		DisclosureWording:   "Werbung",
+	if err != nil {
-		DisclosureBeforeCut: false,
+		t.Fatalf("NewServer: %v", err)
-	}})
+	}
 
-	w := postCheck(t, s, url.Values{"platform": {"instagram"}, "caption": {"Werbung, schaut mal..."}})
+	w := postForm(t, s, "/pruefen", checkForm())
 	if w.Code != http.StatusOK {
-		t.Fatalf("status = %d, want 200", w.Code)
+		t.Fatalf("status = %d, want 200, body: %s", w.Code, w.Body.String())
 	}
-	body := w.Body.String()
+	if !strings.Contains(w.Body.String(), "WK-004") {
-	if !strings.Contains(body, "WK-004") {
+		t.Fatalf("expected WK-004 in result, got: %s", w.Body.String())
-		t.Errorf("expected WK-004 finding in result, got: %s", body)
+	}
+
+	if len(fs.submissions) != 1 {
+		t.Fatalf("expected 1 persisted submission, got %d", len(fs.submissions))
+	}
+	var subID string
+	for id, sub := range fs.submissions {
+		subID = id
+		if sub.Status != "checked" {
+			t.Errorf("submission status = %q, want checked", sub.Status)
+		}
+	}
+	if _, ok := fs.extractions[subID]; !ok {
+		t.Error("expected an extraction to be persisted")
+	}
+	if len(fs.findings[subID]) != 1 || fs.findings[subID][0].RuleID != "WK-004" {
+		t.Errorf("expected exactly one persisted WK-004 finding, got %+v", fs.findings[subID])
 	}
 }
 
-func TestHandleCheckRendersNeedsClarification(t *testing.T) {
+func TestHandleCheckRendersNeedsClarificationAndDoesNotOfferArchive(t *testing.T) {
-	s := newTestServer(t, fakeExtractor{facts: rules.Facts{
+	fs := newFakeStore()
-		Platform:      "instagram",
+	s, err := web.NewServer(fakeExtractor{facts: rules.Facts{
-		Jurisdiction:  "DE",
+		Platform: "instagram", Jurisdiction: "DE", Consideration: rules.ConsiderationUnclear,
-		Consideration: rules.ConsiderationUnclear,
+	}}, loadRealRules(t), fs, fakeTimestamper{}, t.TempDir())
-	}})
+	if err != nil {
-
+		t.Fatalf("NewServer: %v", err)
-	w := postCheck(t, s, url.Values{"platform": {"instagram"}, "caption": {"..."}})
-	if w.Code != http.StatusOK {
-		t.Fatalf("status = %d, want 200", w.Code)
 	}
+
+	w := postForm(t, s, "/pruefen", checkForm())
 	body := w.Body.String()
 	if strings.Contains(body, "WK-") {
 		t.Errorf("expected no rule findings for an unclear extraction, got: %s", body)
@@ -154,16 +370,87 @@ func TestHandleCheckRendersNeedsClarification(t *testing.T) {
 	if !strings.Contains(body, "nicht sicher bestimmt") {
 		t.Errorf("expected a clarification message, got: %s", body)
 	}
+	if strings.Contains(body, "veroeffentlichen") || strings.Contains(body, "/veroeffentlichen") {
+		t.Errorf("expected no archive option when clarification is needed, got: %s", body)
+	}
+	for _, sub := range fs.submissions {
+		if sub.Status == "checked" {
+			t.Error("submission should not be marked checked when clarification is needed")
+		}
+	}
 }
 
-func TestHandleCheckCleanCaseHasNoFindings(t *testing.T) {
+func TestFullCheckThenArchiveFlow(t *testing.T) {
+	fs := newFakeStore()
+	s, err := web.NewServer(fakeExtractor{facts: rules.Facts{
+		Platform: "instagram", Jurisdiction: "DE", Consideration: rules.ConsiderationPaid,
+		DisclosurePresent: true, DisclosureWording: "Werbung", DisclosureBeforeCut: false,
+	}, raw: []byte(`{"gegenleistung":"bezahlt","kennzeichnung_vorhanden":true,"kennzeichnung_wortlaut":"Werbung","kennzeichnung_vor_kuerzung":false}`)},
+		loadRealRules(t), fs, fakeTimestamper{}, t.TempDir())
+	if err != nil {
+		t.Fatalf("NewServer: %v", err)
+	}
+
+	checkResp := postForm(t, s, "/pruefen", checkForm())
+	if checkResp.Code != http.StatusOK {
+		t.Fatalf("check status = %d, body: %s", checkResp.Code, checkResp.Body.String())
+	}
+	if !strings.Contains(checkResp.Body.String(), `name="submission_id"`) {
+		t.Fatalf("expected a hidden submission_id field in the result, got: %s", checkResp.Body.String())
+	}
+
+	var subID string
+	for id := range fs.submissions {
+		subID = id
+	}
+
+	archiveResp := postForm(t, s, "/veroeffentlichen", url.Values{"submission_id": {subID}})
+	if archiveResp.Code != http.StatusOK {
+		t.Fatalf("archive status = %d, body: %s", archiveResp.Code, archiveResp.Body.String())
+	}
+	if !strings.Contains(archiveResp.Body.String(), "/dossier/"+subID) {
+		t.Fatalf("expected a dossier download link, got: %s", archiveResp.Body.String())
+	}
+
+	pkg, ok := fs.evidencePkgs[subID]
+	if !ok {
+		t.Fatal("expected an evidence package to be persisted")
+	}
+	if fs.submissions[subID].Status != "published" {
+		t.Errorf("submission status = %q, want published", fs.submissions[subID].Status)
+	}
+
+	downloadReq := httptest.NewRequest(http.MethodGet, "/dossier/"+subID, nil)
+	downloadReq.SetPathValue("id", subID)
+	downloadW := httptest.NewRecorder()
+	s.ServeHTTP(downloadW, downloadReq)
+	if downloadW.Code != http.StatusOK {
+		t.Fatalf("download status = %d", downloadW.Code)
+	}
+	if !strings.HasPrefix(downloadW.Body.String(), "%PDF-") {
+		t.Fatal("downloaded dossier does not start with the PDF header")
+	}
+
+	if _, err := os.Stat(pkg.DossierPath); err != nil {
+		t.Fatalf("dossier file does not exist on disk: %v", err)
+	}
+}
+
+func TestHandleArchiveRejectsUnknownSubmission(t *testing.T) {
+	s := newTestServer(t, fakeExtractor{})
+
+	w := postForm(t, s, "/veroeffentlichen", url.Values{"submission_id": {"does-not-exist"}})
+	if w.Code != http.StatusNotFound {
+		t.Fatalf("status = %d, want 404 for an unknown submission", w.Code)
+	}
+}
+
+func TestHandleCheckCleanCaseHasNoFindingsAndOffersArchive(t *testing.T) {
 	s := newTestServer(t, fakeExtractor{facts: rules.Facts{
-		Platform:      "instagram",
+		Platform: "instagram", Jurisdiction: "DE", Consideration: rules.ConsiderationNone,
-		Jurisdiction:  "DE",
-		Consideration: rules.ConsiderationNone,
 	}})
 
-	w := postCheck(t, s, url.Values{"platform": {"instagram"}, "caption": {"ein ganz normaler Post"}})
+	w := postForm(t, s, "/pruefen", checkForm())
 	if w.Code != http.StatusOK {
 		t.Fatalf("status = %d, want 200", w.Code)
 	}
@@ -174,10 +461,7 @@ func TestHandleCheckCleanCaseHasNoFindings(t *testing.T) {
 	if !strings.Contains(body, "Keine Kennzeichnungsrisiken") {
 		t.Errorf("expected the no-findings message, got: %s", body)
 	}
+	if !strings.Contains(body, "/veroeffentlichen") {
+		t.Errorf("expected an archive option even with no findings (still a real check), got: %s", body)
+	}
 }
-
-var errTest = extractError("simulierter Extraktionsfehler")
-
-type extractError string
-
-func (e extractError) Error() string { return string(e) }

internal/web/templates/archived.html

@@ -0,0 +1,13 @@
+{{define "archived"}}
+<p class="archiviert">
+  Beitrag archiviert und mit RFC-3161-Zeitstempel versehen
+  ({{.TimestampedAt}}).
+</p>
+<p>
+  <a href="{{.DossierURL}}">Nachweis-Dossier (PDF) herunterladen</a>
+</p>
+<p class="disclaimer">
+  Dieses Dossier dokumentiert die durchgeführte Prüfung. Es ist keine
+  Rechtsberatung und ersetzt keine anwaltliche Prüfung im Einzelfall.
+</p>
+{{end}}

internal/web/templates/index.html

@@ -9,6 +9,14 @@
     <option value="tiktok">TikTok</option>
   </select>
 
+  <label for="post_type">Beitragstyp</label>
+  <select id="post_type" name="post_type" required>
+    <option value="feed">Feed</option>
+    <option value="reel">Reel</option>
+    <option value="story">Story</option>
+    <option value="video">Video</option>
+  </select>
+
   <label for="caption">Caption</label>
   <textarea id="caption" name="caption" rows="6" required></textarea>
 

internal/web/templates/result.html

@@ -20,6 +20,14 @@
 {{else}}
   <p class="keine-findings">Keine Kennzeichnungsrisiken nach aktuellem Regelwerk gefunden.</p>
 {{end}}
+
+{{if .CanArchive}}
+<form hx-post="/veroeffentlichen" hx-target="#ergebnis" hx-swap="innerHTML">
+  <input type="hidden" name="submission_id" value="{{.SubmissionID}}">
+  <button type="submit">Als veröffentlicht markieren &amp; archivieren</button>
+</form>
+{{end}}
+
 <p class="disclaimer">
   Diese Prüfung ist keine Rechtsberatung und ersetzt keine anwaltliche
   Prüfung im Einzelfall.

packaging/etc/deklarix/deklarix.env.example

@@ -9,6 +9,15 @@ PORT=8080
 # /usr/share/deklarix/rules installiert (siehe scripts/build.sh).
 RULES_DIR=/usr/share/deklarix/rules
 
+# Wo erzeugte Nachweis-Dossiers (PDF) abgelegt werden. Wird beim
+# Archivieren eines Beitrags automatisch angelegt.
+DOSSIER_DIR=/var/lib/deklarix/dossiers
+
+# RFC-3161-Zeitstempeldienst. Leer = FreeTSA.org (frei, aber NICHT
+# eIDAS-qualifiziert — siehe CLAUDE.md, Offene Punkte). Vor echtem
+# Kundeneinsatz auf einen eIDAS-qualifizierten Dienst umstellen.
+#TSA_URL=https://freetsa.org/tsr
+
 # Pflicht — der Dienst startet nicht ohne gültige DATABASE_URL und
 # ANTHROPIC_API_KEY. Beide auskommentiert lassen, bis echte Werte
 # eingetragen sind: postinst prüft genau diese beiden Zeilen, um den