diff --git a/include/dynarmic/A32/a32.h b/include/dynarmic/A32/a32.h
index 88f3cf9c..9dab9dc1 100644
--- a/include/dynarmic/A32/a32.h
+++ b/include/dynarmic/A32/a32.h
@@ -35,6 +35,12 @@ public:
      */
     void Run();
 
+    /**
+     * Steps the emulated CPU.
+     * Cannot be recursively called.
+     */
+    void Step();
+
     /**
      * Clears the code cache of all compiled code.
      * Can be called at any time. Halts execution if called within a callback.
diff --git a/src/backend/x64/a32_interface.cpp b/src/backend/x64/a32_interface.cpp
index 8041ff19..18b04e81 100644
--- a/src/backend/x64/a32_interface.cpp
+++ b/src/backend/x64/a32_interface.cpp
@@ -41,7 +41,7 @@ static RunCodeCallbacks GenRunCodeCallbacks(A32::UserCallbacks* cb, CodePtr (*Lo
 
 struct Jit::Impl {
     Impl(Jit* jit, A32::UserConfig config)
-            : block_of_code(GenRunCodeCallbacks(config.callbacks, &GetCurrentBlock, this), JitStateInfo{jit_state})
+            : block_of_code(GenRunCodeCallbacks(config.callbacks, &GetCurrentBlockThunk, this), JitStateInfo{jit_state})
             , emitter(block_of_code, config, jit)
             , config(std::move(config))
             , jit_interface(jit)
@@ -59,13 +59,22 @@ struct Jit::Impl {
     bool invalidate_entire_cache = false;
 
     void Execute() {
-        const u32 new_rsb_ptr = (jit_state.rsb_ptr - 1) & A32JitState::RSBPtrMask;
-        if (jit_state.GetUniqueHash() == jit_state.rsb_location_descriptors[new_rsb_ptr]) {
-            jit_state.rsb_ptr = new_rsb_ptr;
-            block_of_code.RunCodeFrom(&jit_state, reinterpret_cast<CodePtr>(jit_state.rsb_codeptrs[new_rsb_ptr]));
-        } else {
-            block_of_code.RunCode(&jit_state);
-        }
+        const CodePtr current_codeptr = [this]{
+            // RSB optimization
+            const u32 new_rsb_ptr = (jit_state.rsb_ptr - 1) & A32JitState::RSBPtrMask;
+            if (jit_state.GetUniqueHash() == jit_state.rsb_location_descriptors[new_rsb_ptr]) {
+                jit_state.rsb_ptr = new_rsb_ptr;
+                return reinterpret_cast<CodePtr>(jit_state.rsb_codeptrs[new_rsb_ptr]);
+            }
+
+            return GetCurrentBlock();
+        }();
+
+        block_of_code.RunCodeFrom(&jit_state, current_codeptr);
+    }
+
+    void Step() {
+        block_of_code.StepCode(&jit_state, GetCurrentSingleStep());
     }
 
     std::string Disassemble(const IR::LocationDescriptor& descriptor) {
@@ -109,16 +118,21 @@ struct Jit::Impl {
 private:
     Jit* jit_interface;
 
-    static CodePtr GetCurrentBlock(void* this_voidptr) {
+    static CodePtr GetCurrentBlockThunk(void* this_voidptr) {
         Jit::Impl& this_ = *static_cast<Jit::Impl*>(this_voidptr);
-        A32JitState& jit_state = this_.jit_state;
+        return this_.GetCurrentBlock();
+    }
 
-        u32 pc = jit_state.Reg[15];
-        A32::PSR cpsr{jit_state.Cpsr()};
-        A32::FPSCR fpscr{jit_state.upper_location_descriptor};
-        A32::LocationDescriptor descriptor{pc, cpsr, fpscr};
+    IR::LocationDescriptor GetCurrentLocation() const {
+        return IR::LocationDescriptor{jit_state.GetUniqueHash()};
+    }
 
-        return this_.GetBasicBlock(descriptor).entrypoint;
+    CodePtr GetCurrentBlock() {
+        return GetBasicBlock(GetCurrentLocation()).entrypoint;
+    }
+
+    CodePtr GetCurrentSingleStep() {
+        return GetBasicBlock(A32::LocationDescriptor{GetCurrentLocation()}.SetSingleStepping(true)).entrypoint;
     }
 
     A32EmitX64::BlockDescriptor GetBasicBlock(IR::LocationDescriptor descriptor) {
@@ -159,6 +173,18 @@ void Jit::Run() {
     impl->PerformCacheInvalidation();
 }
 
+void Jit::Step() {
+    ASSERT(!is_executing);
+    is_executing = true;
+    SCOPE_EXIT { this->is_executing = false; };
+
+    impl->jit_state.halt_requested = true;
+
+    impl->Step();
+
+    impl->PerformCacheInvalidation();
+}
+
 void Jit::ClearCache() {
     impl->invalidate_entire_cache = true;
     impl->RequestCacheInvalidation();
diff --git a/src/frontend/A32/location_descriptor.cpp b/src/frontend/A32/location_descriptor.cpp
index 08391a27..0c20247e 100644
--- a/src/frontend/A32/location_descriptor.cpp
+++ b/src/frontend/A32/location_descriptor.cpp
@@ -11,11 +11,12 @@
 namespace Dynarmic::A32 {
 
 std::ostream& operator<<(std::ostream& o, const LocationDescriptor& descriptor) {
-    o << fmt::format("{{{:08x},{},{},{:08x}}}",
+    o << fmt::format("{{{:08x},{},{},{:08x}{}}}",
                      descriptor.PC(),
                      descriptor.TFlag() ? "T" : "!T",
                      descriptor.EFlag() ? "E" : "!E",
-                     descriptor.FPSCR().Value());
+                     descriptor.FPSCR().Value(),
+                     descriptor.SingleStepping() ? ",step" : "");
     return o;
 }
 
diff --git a/src/frontend/A32/location_descriptor.h b/src/frontend/A32/location_descriptor.h
index 80d38cff..78c056d4 100644
--- a/src/frontend/A32/location_descriptor.h
+++ b/src/frontend/A32/location_descriptor.h
@@ -29,8 +29,12 @@ public:
     static constexpr u32 CPSR_MODE_MASK  = 0x0600FE20;
     static constexpr u32 FPSCR_MODE_MASK = 0x07F70000;
 
-    LocationDescriptor(u32 arm_pc, PSR cpsr, FPSCR fpscr)
-            : arm_pc(arm_pc), cpsr(cpsr.Value() & CPSR_MODE_MASK), fpscr(fpscr.Value() & FPSCR_MODE_MASK) {}
+    LocationDescriptor(u32 arm_pc, PSR cpsr, FPSCR fpscr, bool single_stepping = false)
+        : arm_pc(arm_pc)
+        , cpsr(cpsr.Value() & CPSR_MODE_MASK)
+        , fpscr(fpscr.Value() & FPSCR_MODE_MASK)
+        , single_stepping(single_stepping)
+    {}
 
     explicit LocationDescriptor(const IR::LocationDescriptor& o) {
         arm_pc = static_cast<u32>(o.Value());
@@ -38,6 +42,7 @@ public:
         cpsr.E((o.Value() >> 32) & 2);
         fpscr = (o.Value() >> 32) & FPSCR_MODE_MASK;
         cpsr.IT(ITState{static_cast<u8>(o.Value() >> 40)});
+        single_stepping = (o.Value() >> 32) & 4;
     }
 
     u32 PC() const { return arm_pc; }
@@ -48,8 +53,10 @@ public:
     A32::PSR CPSR() const { return cpsr; }
     A32::FPSCR FPSCR() const { return fpscr; }
 
+    bool SingleStepping() const { return single_stepping; }
+
     bool operator == (const LocationDescriptor& o) const {
-        return std::tie(arm_pc, cpsr, fpscr) == std::tie(o.arm_pc, o.cpsr, o.fpscr);
+        return std::tie(arm_pc, cpsr, fpscr, single_stepping) == std::tie(o.arm_pc, o.cpsr, o.fpscr, single_stepping);
     }
 
     bool operator != (const LocationDescriptor& o) const {
@@ -57,36 +64,40 @@ public:
     }
 
     LocationDescriptor SetPC(u32 new_arm_pc) const {
-        return LocationDescriptor(new_arm_pc, cpsr, fpscr);
+        return LocationDescriptor(new_arm_pc, cpsr, fpscr, single_stepping);
     }
 
     LocationDescriptor AdvancePC(int amount) const {
-        return LocationDescriptor(static_cast<u32>(arm_pc + amount), cpsr, fpscr);
+        return LocationDescriptor(static_cast<u32>(arm_pc + amount), cpsr, fpscr, single_stepping);
     }
 
     LocationDescriptor SetTFlag(bool new_tflag) const {
         PSR new_cpsr = cpsr;
         new_cpsr.T(new_tflag);
 
-        return LocationDescriptor(arm_pc, new_cpsr, fpscr);
+        return LocationDescriptor(arm_pc, new_cpsr, fpscr, single_stepping);
     }
 
     LocationDescriptor SetEFlag(bool new_eflag) const {
         PSR new_cpsr = cpsr;
         new_cpsr.E(new_eflag);
 
-        return LocationDescriptor(arm_pc, new_cpsr, fpscr);
+        return LocationDescriptor(arm_pc, new_cpsr, fpscr, single_stepping);
     }
 
     LocationDescriptor SetFPSCR(u32 new_fpscr) const {
-        return LocationDescriptor(arm_pc, cpsr, A32::FPSCR{new_fpscr & FPSCR_MODE_MASK});
+        return LocationDescriptor(arm_pc, cpsr, A32::FPSCR{new_fpscr & FPSCR_MODE_MASK}, single_stepping);
     }
 
     LocationDescriptor AdvanceIT() const {
         PSR new_cpsr = cpsr;
         new_cpsr.IT(new_cpsr.IT().Advance());
 
-        return LocationDescriptor(arm_pc, new_cpsr, fpscr);
+        return LocationDescriptor(arm_pc, new_cpsr, fpscr, single_stepping);
+    }
+
+    LocationDescriptor SetSingleStepping(bool new_single_stepping) const {
+        return LocationDescriptor(arm_pc, cpsr, fpscr, new_single_stepping);
     }
 
     u64 UniqueHash() const noexcept {
@@ -96,8 +107,9 @@ public:
         const u64 fpscr_u64 = fpscr.Value();
         const u64 t_u64 = cpsr.T() ? 1 : 0;
         const u64 e_u64 = cpsr.E() ? 2 : 0;
+        const u64 single_stepping_u64 = single_stepping ? 4 : 0;
         const u64 it_u64 = u64(cpsr.IT().Value()) << 8;
-        const u64 upper = (fpscr_u64 | t_u64 | e_u64 | it_u64) << 32;
+        const u64 upper = (fpscr_u64 | t_u64 | e_u64 | single_stepping_u64 | it_u64) << 32;
         return pc_u64 | upper;
     }
 
@@ -109,6 +121,7 @@ private:
     u32 arm_pc;       ///< Current program counter value.
     PSR cpsr;         ///< Current program status register.
     A32::FPSCR fpscr; ///< Floating point status control register.
+    bool single_stepping;
 };
 
 /**
diff --git a/src/frontend/A32/translate/translate_arm.cpp b/src/frontend/A32/translate/translate_arm.cpp
index a0d29b36..8dd85983 100644
--- a/src/frontend/A32/translate/translate_arm.cpp
+++ b/src/frontend/A32/translate/translate_arm.cpp
@@ -33,8 +33,9 @@ IR::Block TranslateArm(LocationDescriptor descriptor, MemoryReadCodeFuncType mem
     IR::Block block{descriptor};
     ArmTranslatorVisitor visitor{block, descriptor, options};
 
+    const bool single_step = descriptor.SingleStepping();
     bool should_continue = true;
-    while (should_continue && CondCanContinue(visitor.cond_state, visitor.ir)) {
+    do {
         const u32 arm_pc = visitor.ir.current_location.PC();
         const u32 arm_instruction = memory_read_code(arm_pc);
 
@@ -52,11 +53,15 @@ IR::Block TranslateArm(LocationDescriptor descriptor, MemoryReadCodeFuncType mem
 
         visitor.ir.current_location = visitor.ir.current_location.AdvancePC(4);
         block.CycleCount()++;
-    }
+    } while (should_continue && CondCanContinue(visitor.cond_state, visitor.ir) && !single_step);
 
-    if (visitor.cond_state == ConditionalState::Translating || visitor.cond_state == ConditionalState::Trailing) {
+    if (visitor.cond_state == ConditionalState::Translating || visitor.cond_state == ConditionalState::Trailing || single_step) {
         if (should_continue) {
-            visitor.ir.SetTerm(IR::Term::LinkBlockFast{visitor.ir.current_location});
+            if (single_step) {
+                visitor.ir.SetTerm(IR::Term::LinkBlock{visitor.ir.current_location});
+            } else {
+                visitor.ir.SetTerm(IR::Term::LinkBlockFast{visitor.ir.current_location});
+            }
         }
     }
 
diff --git a/src/frontend/A32/translate/translate_thumb.cpp b/src/frontend/A32/translate/translate_thumb.cpp
index 21b0949c..0dd6c1c9 100644
--- a/src/frontend/A32/translate/translate_thumb.cpp
+++ b/src/frontend/A32/translate/translate_thumb.cpp
@@ -59,8 +59,9 @@ IR::Block TranslateThumb(LocationDescriptor descriptor, MemoryReadCodeFuncType m
     IR::Block block{descriptor};
     ThumbTranslatorVisitor visitor{block, descriptor, options};
 
+    const bool single_step = descriptor.SingleStepping();
     bool should_continue = true;
-    while (should_continue) {
+    do {
         const u32 arm_pc = visitor.ir.current_location.PC();
         const auto [thumb_instruction, inst_size] = ReadThumbInstruction(arm_pc, memory_read_code);
 
@@ -81,6 +82,10 @@ IR::Block TranslateThumb(LocationDescriptor descriptor, MemoryReadCodeFuncType m
         const s32 advance_pc = (inst_size == ThumbInstSize::Thumb16) ? 2 : 4;
         visitor.ir.current_location = visitor.ir.current_location.AdvancePC(advance_pc);
         block.CycleCount()++;
+    } while (should_continue && !single_step);
+
+    if (single_step && should_continue) {
+        visitor.ir.SetTerm(IR::Term::LinkBlock{visitor.ir.current_location});
     }
 
     block.SetEndLocation(visitor.ir.current_location);